Producing nitrogen (N) uptake maps in winter wheat by combining proximal crop measurements with Sentinel-2 and DMC satellite images in a decision support system for farmers

Abstract

Responsive fertilisation of winter wheat (Triticum aestivum L.) is often adopted, with N applied two or three times between the developmental stages of tillering and booting. Satellite-based decision support systems (DSS) providing vegetation index maps calculated from satellite data are available to aid farmers in adjusting the topdressing nitrogen (N) rate site-specifically to the current season and to variations in growth conditions within the field. One example is the freely available CropSAT DSS used in Scandinavia, which provides farmers with raster maps of the modified soil-adjusted vegetation index (MSAVI2) calculated mainly from data obtained from satellites Sentinel-2 (ESA, EU) and DMC (DMCii Ltd, Guildford, UK). This study investigated the possibility of calibrating MSAVI2 maps with data from handheld proximal sensor measurements of N uptake covering the main agricultural regions in Sweden during growth stages Z30-45 on the Zadok scale, in order to facilitate farmers’ decisions on N rate. More than 200 N-sensor measurements acquired during 2015 and 2016 in seven different winter wheat cultivars were combined with MSAVI2 values from CropSAT. It was found that N uptake could be predicted in a general, national model, i.e. for sites and dates other than those for which the calibration model was parameterised, with a mean absolute error of 11–15 kg N ha−1. A cultivar-specific model performed better than this general model, but a regional model showed no improvement compared with the model parameterised with national data. Vegetation indices calculated from the two narrow bands of Sentinel-2 in the red edge-near infrared region of the crop canopy reflectance spectrum proved to be promising alternatives to the broadband index MSAVI2. Based on the results, we suggest that data from a monitoring programme involving handheld N sensor measurements can be integrated with a satellite-based DSS to upscale N uptake information.

Keywords

• nitrogen fertilisation
• decision support system
• remote sensing
• proximal sensing
• yara n-sensor
• upscaling
• cropsat
• variable-rate application
• precision agriculture