Evaluating the water status of winter wheat using ground-based spectral data

Abstract

Abstract Tracking crop water status is important for assessing crop water balance and developing water-saving irrigation strategies. These actions are of great theoretical and practical significance for promoting sustainable use of regional water resources and for improving crop water use efficiency. We conducted experimental field trials in 2012–2016 for winter wheat (Triticum aestivum L.) under three water treatments (rainfed, deficit-irrigated, and fully-irrigated). Canopy spectral reflectance and leaf water content were measured during the growing season. The Normalized Difference Vegetation Index (NDVI), Ratio Vegetation Index (RVI), and Difference Vegetation Index (DVI) were calculated using all possible combinations of two spectral reflectance bands between 451 nm and 2400 nm. Correlations between these vegetation indices (VIs) and leaf water content before and after irrigation were evaluated. Finally, we established estimation models of leaf water content and compared 16 commonly used VIs (such as NDII, WI, and WBI) at 144 trial plots to select the optimal vegetation index and wavebands. We found that leaf water content and VIs for the three water treatments followed the order of fully-irrigated > deficit-irrigated > and rainfed. Leaf spectral reflectance increased from greening to jointing, and then decreased from jointing to harvest. The spectral bands that were sensitive to crop water content were mainly observed in the visible and near-infrared regions. The highest correlation between leaf water content and VIs was for NDVI when using spectral bands at 1191 nm and 1305 nm. A predictive model was subsequently proposed that accounted for 82% of the leaf water content variation. The average R-square for all VIs was 0.80, indicating that a number of uncertainties remain when only using VIs to track irrigation activity. These results provide guidance for selecting spectral bands when developing portable instruments for monitoring crop water status. Our method to monitor crop water status and irrigation activities is a template that can be used at regional scales.