Monitoring winter wheat growth performance at sub-field scale using multitemporal Sentinel-2 imagery

Abstract

A crop growth monitoring system should objectively and reproducibly reflect changes in crop biophysical properties during the growing season. By monitoring crop growth and performance at specific crop development stages, the farmer can obtain reliable information for timely crop management to achieve optimum crop production. This work aimed to evaluate crop development using five winter wheat (Triticum aestivum L.) biophysical properties (shoots number, green area index, plant height, leaf N content, and aboveground dry biomass) predicted from Sentinel-2 data compared with benchmarks representing target growth from emergence to harvest. Data were collected for four principal phenology stages (tillering, stem elongation, heading, and fruit development) in 35 winter wheat fields in the Republic of Ireland and 40 in the United Kingdom in 2020 and 2021. A total of 1500 plots were selected for crop sampling over two growing seasons. The models were generally good, but phenology-specific models performed better (R² between 0.72 and 0.87) than models for the entire season (R² between 0.13 and 0.84). To assess the low-performance zones in fields, the predicted biophysical properties were compared to benchmarks taken from agronomic advice. Spatial analysis was then used to identify low-performance areas in fields, which were validated using farmers’ feedback. It was concluded that the approach taken could be reliably used to monitor winter wheat over a wide area and through time.

More information

Identifiers

ORCID for Bing-Bing Goh: orcid.org/0000-0001-7108-991X

Catalogue record

Export record

Altmetrics