The effects of N and P canopy content on the hyperspectral signature of vegetation and their application in remote sensing

Abstract

Remote sensing techniques have been developed for monitoring large areas of vegetation at once. Hyperspectral reflectance spectroscopy allows for detailed study of the reflectance profile of vegetation. Plants require nitrogen and phosphorus for photosynthesis and as components of many organic compounds, for tissue growth, and for energy storage. As a result, N and P availability influences the hyperspectral signature of vegetation, but its effect has not been intensively studied. This study investigates the effects of N and P canopy content on the hyperspectral signature of Holcus lanatus via a greenhouse experiment, and whether the observed trends could be studied using existing remote sensing satellite platforms. It is also assessed how representative these trends are for the hyperspectral response of vegetation in general to changes in nutrient availability via comparison to MTCI and canopy nutrient data from Catalonia. This is supported by a literature review on remote sensing and vegetation indices. The hyperspectral signature of Holcus lanatus is found to be affected mainly between 500 – 1320 nm, but also between 1450 - 1900 nm and 2100 – 2300 nm. High P content is associated with low reflectance values, and high N content is associated with high reflectance values, but their individual contributions could not be accurately determined. Based on the simulated performance of Envisat MERIS, Landsat 7 ETM+, and Sentinel 2 MSI, these satellite platforms were able to detect these trends. MTCI and other red-edge based wavelength combinations were found to respond best based on changes in nutrient availability, with an extinction curve MTCI response to increasing P concentrations and a positive linear MTCI response to increasing N concentrations. NDVI and SAVI were found to have no significant correlation to either canopy P or N content. Other types of flora from the Catalonian dataset have a positive linear MTCI response to both increasing P and N concentrations, showing MTCI a valuable tool in assessing canopy N content. The difference in P response emphasizes the importance of knowing the type of vegetation in an area and its specific spectral response to changes in nutrients before assessing them using remote sensing techniques.