Comparing optical and direct methods for leaf area index determination in a maize crop, Journal of Agricultural Engineering, 1, 27-34

Leaf area index (LAI) is a crucial variable in the modelling of many hydrological processes. Destructive sampling of LAI is extremely time-consuming, thus not suitable for monitoring temporal/spatial variations of the variable. In the last fifty years optical instruments retrieving LAI from more easily measurable variables (i.e. transmitted radiation through canopies) have been developed. Several instruments are available on the market, but very few are the studies comparing LAI estimates in agricultural crops. In this paper three optical instruments are compared with destructive sampling for a maize crop located in Northern Italy. Determinations were carried out on three plots (replicates) before and after successive thinning of plant populations. Destructively sampled LAI ranged from 4.9 m2m-2 (no thinning) to 1.2 m2m-2 (maximum thinning). Correlation analysis showed that estimates by the AccuPAR-80, the hemispherical camera (“effective” and “true” LAI) and the LAI-2000 (in the standard configuration, i.e. five zenithal rings, and excluding the fifth ring) were well correlated with destructive measurements (R2>0.95). Anyway, if for the AccuPAR-80 the regression line was close to the 1:1 line, the “true LAI” by hemispherical photography tended to overestimate destructively sampled LAI for low values while the “effective LAI” and the LAI-2000 to underestimate it for high values (in a minor way for the LAI-2000 when the fifth ring was removed). Results of ANOVA and Tukey T test for two-factor experiments with replicates showed that only the “effective LAI” retrieved by hemispherical photographs and the estimates provided by the LAI-2000 in the standard configuration (five rings) were statistically different from destructive measurements.