Reducing the error in biomass estimates strongly depends on model selection

Table 1 Allometric equations to assess the error due to model choice

Reference	Country	No.	Expression	σ	n	\(D_{\min }\)	\(D_{\max }\)
Chave et al. (2005)	Pantropical	II.3	\(\rho \exp \left [-1.499+2.148\ln D+\right .\)	0.356 ^a	1504	5	156
			\(\left .0.207(\ln D)^{2}-0.0281(\ln D)^{3}\right ]\)
Chave et al. (2005)	Pantropical	I.5	0.0509×ρ D ² H	0.316 ^a	1349	5	156
Dorisca et al. (2011)	Cameroon	5	\(\rho \times \left (11.1956+0.040326D^{2}H\right )\)	0.374 ^b	51	10	134
Dorisca et al. (2011)	Cameroon	2	\(\rho \times \left (325.5573-47.2984D+\right .\)	0.254 ^c	55	10	134
			\(\left .2.2942D^{2}\right )\)
Fayolle et al. (2013)	Cameroon	2	\(\rho \exp \left [-1.183+1.940\ln D+\right .\)	0.188 ^a	138	5	192
			\(\left .0.239(\ln D)^{2}-0.0285(\ln D)^{3}\right ]\)
Ngomanda et al. (2014)	Gabon	5	\(\exp \left [-4.060+4.062\ln D-\right .\)	0.330 ^a	101	12	109
			\(\left .0.228(\ln D)^{2}+1.431\ln \rho \right ]\)
Ngomanda et al. (2014)	Gabon	11	\(\exp \left [-2.568+0.952\ln (D^{2}H)+\right .\)	0.292 ^a	101	12	109
			\(\left .1.189\ln \rho \vphantom {D^{2}}\right ]\)

In all these equations, the predicted quantity is the tree aboveground biomass in kg, D is dbh in cm, H is height in m and ρ is wood density in g cm ⁻³. The number (No.) is the model number in the publication. σ is the residual standard error of the model. n is the number of observations used to fit the model. \(D_{\min }\) and \(D_{\max }\) are the minimum and maximum dbh (in cm) in the data set used to fit the model. The expressions of the equations include the correction factor for back-transformation when the model was fitted to log-transformed data
^aResidual of log-transformed data
^bThe standard deviation of residuals is σ D ^2.2
^cThe standard deviation of residuals is σ D ^2.15

ISSN: 1297-966X