Table II

Comparison of the accuracy and the robustness of 4 competitive general models developed to predict chlorophyll content (Chl) with a SPAD-502. Accuracy was estimated using the Akaike Information Criterion (AIC). Robustness was estimated using both the coefficient of determination R² and the Root Mean Square Error of Prediction (RMSEP) throughout a cross-validation procedure.

Model	α	β	AIC		Min R2	Mean R2	Max R2			Min RMSEP	Mean RMSEP	Max RMSEP
Linear
Chli = α * SPADi	1.310 ± 0.009	_	3215.9		0.8057	0.8454	0.8868			157.7	207.2	251.3
Polynomial
	0.664 ± 0.040	0.012 ± 0.001	3075.3		0.8485	0.8794	0.9174			141.4	175.0	208.8
Exponential
Chli = α(eβ * SPADi − 1)	0.018 ± 0.001	42.85 ± 4.48	3049.2		0.8554	0.8853	0.9201			139.1	170.7	204.3
Homographic
	117.10 ± 6.53	148.84 ± 8.69	3029.7		0.8598	0.8899	0.9211			137.3	164.9	201.4