Evapotranspiration of winter wheat estimated with the FAO 56 approach and NDVI measurements in a temperate humid climate of NW Europe

Abstract Irrigated agriculture is the dominant user of global freshwater resources and improvements in water management are considered key to simultaneously enhance food production and to secure regional water resources. Irrigation management requires reliable estimates of crop evapotranspiration (ET a ) which can be derived from tabulated crop coefficients and estimates of phenological stage lengths or alternatively from remotely sensed canopy data. In order to compare both approaches under sufficient nitrogen and water supply, field trials with winter wheat ( Triticum aestivum L.) were performed in 2013/14 and 2014/15 at Duelmen, Germany. Crop growth dynamics were measured by a handspectrometer and leaf canopy analyzer, and phenological stages were recorded. ET a was calculated in two different ways: Firstly based on tabulated crop coefficients (K c ) and stage durations (FA-approach) and secondly based on remotely sensed K c -values (NDVI-approach). Both approaches were used to calculate a daily field water balance (FWB) and compared with measured (FDR-techniques) dynamics of soil water content. Both ET a estimates predicted seasonal dynamics of the FWB reasonably good RMSE: 0.7, 0.6 (FA-approach); 0.6, 0.8 (NDVI-approach) in 2014 and 2015, respectively. ET a estimates of the FA approach for 2014 and 2015 were 403 mm and 430 mm and exhibited larger between-year variation when the NDVI approach was used (377 mm and 463 mm). Agronomic water-use efficiencies varied between 1.93–2.76 g/l. The relevance of reliable in-field estimates of Kc-values and stage lengths was quantified with a sensitivity analysis. The results indicate that NDVI-based estimates of ET a can be used instead of ET a estimates which are based on tabulated Kc values and estimates of growth stage length. The use of NDVI-based estimates of ET a enables to apply irrigation water more site-specific.