Vertical Chlorophyll Canopy Structure Affects the Remote Sensing Based Predictability of LAI, Chlorophyll and Leaf Nitrogen in Agricultural Fields

Leaf nitrogen and leaf surface area influence the exchange of gases between terrestrial ecosystems and the atmosphere, and they play a significant role in the global cycles of carbon, nitrogen and water. Remote sensing can be used to estimate leaf area index (LAI), chlorophyll content (CHL) and leaf nitrogen (N), but methods are often developed using plot-scale data and not verified over extended regions characterized by variations in environmental boundary conditions (soil, atmosphere) and canopy structures. Estimation of N can be indirect due to its association with CHL, however N is also included in pigments such as carotenoids and anthocyanin which have different spectral signatures than CHL. Photosynthesis optimization theory suggests that plants will distribute their N resources in proportion to the light gradient within the canopy. Such vertical variation in CHL and N complicates the evaluation of remote sensing-based methods. Typically remote sensing studies measure CHL of the upper leaf, which is then multiplied by the green LAI to represent canopy chlorophyll content, or random sampling is used. In this study, field measurements and high spatial resolution (10-20 m) remote sensing images acquired from the HRG and HRVIR sensors aboard the SPOT satellites were used to assess the predictability of LAI, CHL and N in five European agricultural landscapes located in Denmark, Scotland (United Kingdom), Poland, The Netherlands and Italy . All satellite images were atmospherically using the 6SV1 model with atmospheric inputs estimated by MODIS and AIRS data. Five spectral vegetation indices (SVIs) were calculated (the Normalized Difference Vegetation index, the Simple Ratio, the Enhanced Vegetation Index-2, the Green Normalized Difference Vegetation Index, and the green Chlorophyll Index), and an image-based inverse canopy radiative transfer modelling system, REGFLEC (REGularized canopy reFLECtance) was applied to each of the five European landscapes. While the SVIs require field data for empirical model building, the REGFLEC model was applied without calibration. LAI and SPAD meter data were measured in 93 fields representing 10 crop types of the five European landscapes. SPAD meter data were measured at five canopy height levels and converted to CHL and N using laboratory calibration. The data showed strong vertical leaf chlorophyll gradient profiles in 20 % of fields. This affected the predictability of SVIs and REGFLEC. However, selecting only homogeneous canopies with uniform CHL distributions as reference data for statistical evaluation, significant predictions were achieved for all landscapes, by all methods, with the best overall results given by REGFLEC. Predictabilities of SVIs and REGFLEC simulations improved when constrained to single land use categories across the European landscapes, reflecting sensitivity to canopy structures, and predictabilities further improved when constrained to local (10 x 10 km2) landscapes, thereby reflecting sensitivity to local environmental conditions. The Enhanced Vegetation Index-2 tended to be the best method in landscapes with high vegetation densities, REGFLEC worked best in a landscape with large contrasts in vegetation density, and the Simple Ratio worked best in a landscape characterized by low vegetation density.