The many challenges facing European agriculture and farm of tomorrow are such that they increasingly require the setting up of Decision Support Systems (DSS) that favour integrated crop management at farm or regional level. A valuable DSS for management of split fertilizer N applications was developed in Belgium for potato crop. It combines total N recommendation based on field predictive balance-sheet method along with Crop Nitrogen Status (CNS) monitoring through hand-held chlorophyll meter (CM) Hydro N-tester (Yara, Oslo, Norway) readings. In-season monitoring is realized from 20 to 50 days after potato crop emergence. Using zero N reference plot (10x20m) within supposed homogeneous field areas, relative CM threshold value allows to decide on the need for supplemental N to refine the previous recommendation.

Potentialities of Earth Observation and crop light reflectance data to monitor CNS have been investigated for the growing seasons 2008 and 2009, on plots of different sizes (ranging from 30x30m up to 100x100m) fertilized with various mineral N-rates (0, 70 % and 100 % of N recommendation). Reflectances were acquired from SPOT 5 multispectral images (10 m pixel resolution, Red/Green/Near-Infra Red/Short-Wave Infra Red) or from ground-based remote sensing with a hand-held radiometer. Both observations type were planed to be realize from mid of June up to end of July.

The analysis of satellite image was realized following two approaches: (i) the use of statistical analyses to determine relevant spectral bands and (ii) the calculation of spectral Vegetation Indices (VI’s). The R and the NIR bands and several VI’s derived from both bands but also from the G band, discriminated well the various N-rates, but particularly the zero-N plots as did CM readings. DN values between pixels of a same plot varied little (cv < 5%), while variations between same N-rate plots of different fields were high. Discrimination between plots of smallest size based on single pixel DN value was also quite effective.

Very good correlations (r2> 0.80) appeared between VI’s based on satellite DN values and radiometer readings. This prospective study indicates that SPOT 5 MS images show promise to monitor CNS using small size reference plots. The main limitation remains the availability of at least one SPOT 5 image within the required phenological period.