Using Hyperspectral Remote Sensing Data for Retrieving Canopy Chlorophyll and Nitrogen Content

Abstract

Plant stress is often expressed as a reduction in amount of biomass or leaf area index (LAI). In addition, stress may affect the plant pigment system, influencing the photosynthetic capacity of plants. Chlorophyll content is the main driver for this primary production. The chlorophyll content is indirectly related to the nitrogen (N) content. In this paper emphasis is on estimation of canopy chlorophyll content and N content using remote sensing techniques. Hyperspectral reflectance data representing a range of canopies were simulated using the PROSAIL radiative transfer model at a 1 nm sampling interval. Various indices were tested for estimating canopy chlorophyll content. Subsequently, tests with field data were performed for sampling locations within an extensively grazed fen meadow using ASD FieldSpec measurements and within a potato field with a Cropscan radiometer for estimating canopy N content. PROSAIL simulations showed that the red-edge chlorophyll index (CI <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">red edge</sub> ) was linearly related to the canopy chlorophyll content over the full range of potential values (R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> =0.94) . In contrast, highly non-linear relationships of chlorophyll content with most traditional red-edge indices were found. At the study sites the CI <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> was found to be a good and linear estimator of canopy N content (no chlorophyll was measured) for both the grassland site (R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> =0.77) and for the potato field (R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> =0.88) . The latter number refers to plots showing no “luxury” N consumption. However, for the full potato data set, including highly fertilized plants, an exponential relationship yielded a better fit (R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> =0.85) as compared to a linear fit (R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> =0.65) . Currently, this approach can, e.g., be applied with MERIS and Hyperion data and with the upcoming Sentinel-2 and -3 systems.