Do Decision Support Tools Allow Farmers to be Better Advised on Nitrogen Fertilisation in Wheat—Rapeseed Crops Succession in Northern France?

Abstract

The use of sensor- or model-based decision support tools (DSTs) is encouraged to manage crop nitrogen (N) fertilisation, but studies evaluating the obtained advice at the crop succession scale are lacking. The aim of this work was to compare the agronomic performance between the advice obtained by DSTs and that obtained by the reference balance sheet method (BSM) regarding the N fertiliser dose and to evaluate the consequences for soil mineral N (SMN) dynamics. In this investigation, we studied winter wheat/rapeseed/winter wheat crop succession at two experimental sites in northern France according to a randomised block design with four replications. Microplots receiving N fertiliser doses determined by two sensor-based techniques (proximal sensing or PS and remote sensing with a UAV or RS-U) and one model-based DST (dynamic balance model or DBM) were compared to those receiving N fertiliser doses determined by the BSM and to the unfertilised control (T0) plot. In 67% of the studied cases, the N fertiliser dose determined by DSTs was higher than that determined by the BSM, without significant increases in the grain yield and quality: DBM (five out of six cases) > PS (four out of six cases) > RS-U (three out of six cases). The difference in the recommended N fertiliser dose between the DSTs and BSM ranged from −29 to +80 kg N ha⁻¹. The surplus of applied N was underutilised by crops, particularly during dry years, leading to overfertilisation. In our experimental condition, beyond an overfertilisation value of 42 kg N ha⁻¹, a positive relationship was observed between the excess of SMN observed at harvest (R² = 0.39***) and the overfertilisation. Our results indicated high variability of the recommended N fertiliser dose depending on the chosen DSTs, eventually inducing overfertilisation and excess SMN at harvest without significant changes in the grain yield and quality. We consider that improving the analysis of plant or soil N status evolution whilst considering weather parameters in DSTs should improve the accuracy of the provided N fertilisation advice.