Optimal geometric configuration and algorithms for LAI indirect estimates under row canopies: The case of vineyards

Leaf Area Index (LAI) retrieval performances from gap fraction measurements are investigated over vertically trained vineyards. A 3D vineyard model was constructed to analyze the influence of canopy architecture characteristics and light direction on LAI estimation. Results show that for specific directions – close to zenith and parallel to the rows – gap fraction (Po) is mainly driven by vineyard architectural characteristics with small effect of LAI due to the clumped foliage distribution. The sensitivity of Po to LAI is enhanced for directions far from zenith and perpendicular to the rows, resulting in lower uncertainties in LAI retrieval. The 3D vineyard model was used to simulate a range of cases and those findings were supported with two calibrated neural networks to retrieve LAI from Po measured in several directional configurations. These relationships were tested over independent field experiments conducted on commercial vineyards using either hemispherical photographs or ceptometer. Both experiments highlighted the importance of selecting an appropriate geometrical configuration and introducing information on canopy architecture (canopy height and vegetation width on the row relative to inter-row spacing) to reduce LAI uncertainties associated with vineyard spatial arrangement. For the optimal configurations, estimates of LAI with hemispherical photographs (RMSE = 0.389) and ceptometer (RMSE = 0.27) were obtained and compared to destructive LAI measurements.