Correcting Optical Ground Measurement of Leaf Area Index (LAI) Using the Path Length Distribution Model in California Vineyards: Increasing Accuracy for Validating Remote Tools

Leaf Area Index (LAI) is a critical variable to scale satellite based Evapotranspiration (ET) estimations. Indirect LAI measures from satellite observations are in need of ground validation, especially over complex crop surfaces such as vineyards. While direct destructive sampling methods are ideal, there is still need for rapid and accurate indirect ground measurements of LAI. Optical methods for LAI ground retrieval have become a favorable option for ground measurement due to their balance between accuracy and ability to cover large areas. However, optic method calculations assume homogeneity of foliar spatial distribution. Most of the world’s forests and perennial cropping systems have heterogeneous canopy cover. California's winegrape vineyards have vines trained on trellis systems of multiple canopy layers leading to high spatial variability of canopy height and leaf area density. Underestimation of LAI using optic methods has often been attested to clumping conditions, where foliage is heterogeneously distributed, and to occlusion, where foliage is vertically layered. Therefore, novel methods of LAI ground measurement must be explored to account for canopy clumping. The Path Length Distribution Model (PATH) (Hu et al. 2014) uses a spatial distribution of ray path lengths through the canopy from the sensor to correct indirect in-situ LAI against non-uniform canopy height. The Path Measurement Tool (PMT) was developed during this study to measure canopy path length distributions within digital mesh reconstructions at all in-situ gap fraction measurement locations. In this study, we attempt to correct optical in-situ LAI retrievals from a Cabernet Sauvignon vineyard in central California using the PMT and PATH model. To validate the corrections, direct LAI was collected via destructive methods. Gap fraction values were retrieved using the LAI-2200. Path lengths were measured on 3D reconstructions of the measured grapevines sourced from terrestrial LiDAR (TLS) and UAV based photogrammetry. Corrections using path lengths measured from photogrammetry are compared with those measured from TLS to validate a more accessible measurement technique. Path length corrected ground measurements will be used to validate satellite based LAI models. Preliminary results suggest an average increase of 4.39 in LAI (p = 4.6x10^-5) from Beer’s law calculations to path length corrected calculations when using TLS.