Remote sensing based mapping of leaf nitrogen and leaf area index in European landscapes using the REGularized canopy reflectance (REGFLEC) model

Leaf biochemistry and biophysical parameters are important for simulating soil-vegetation-atmosphere exchanges
of energy, water, CO2 and nitrogen. The accumulation of leaf nitrogen (N) in vegetation canopies is a major
component of the ecosystem N balance, and leaf N concentration and leaf area index (LAI) are also important
determinants of the maximum photosynthetic capacity and CO2 uptake by plants and trees. Even though agroand
ecosystem-models can simulate leaf N uptake and LAI of vegetation canopies, regional modeling requires
detailed spatial information about soil properties, N fertilization and atmospheric N deposition rates which are not
easily accessible. In this study, high spatial resolution remote sensing data from the SPOT satellites were acquired
within the NitroEurope project to prepare maps of leaf N and LAI for 5 European landscapes. Mapping was
conducted using the REGFLEC model which is an automatic and image-based methodology recently developed for
regional chlorophyll (Cab) and LAI estimation (ie. Houborg and Anderson, JARS 3, 2009). REGFLEC combines
models for atmospheric correction (6S), canopy reflectance (ACRM) and leaf optics (PROSPECT). The only input
information required are sensor characteristics, atmospheric properties (ie. derived from AIRS and MODIS satellite
sensors) and maps of soil types and spectral vegetation classes within the study area. REGFLEC solves for the
soil background reflectance and builds land cover specific look-up tables which are facilitating fast computation
of Cab and LAI from spectral band reflectances or vegetation indices. Model performance previously proved very
promising in Denmark (ie.Houborg and Boegh, Remote Sens. Env., 112, 2008) and in Maryland, USA (ie. Houborg
et al., Remote Sens. Env., 113, 2009). In this study, REGFLEC performance is evaluated and discussed using field
measurements of leaf N, SPADmeter data (SPAD 502 DL) and LAI (LAI-2000) in European landscapes located in
Denmark, Poland, Scotland, the Netherlands and Italy. The inverse model estimations of soil reflectance parameters
and canopy parameters are discussed in relation to the prevailing soil types and vegetation characteristics of land
cover classes across the 5 European landscapes