Satellite-based NDVI crop coefficients and evapotranspiration with Eddy Covariance Validation for multiple durum wheat fields in the US Southwest

Authors: French, Andrew; Hunsaker, Douglas - Doug; SANCHEZ, CHARLES - University Of Arizona; SABER, MAZIN - University Of Arizona; GONZALEZ, JUAN - University Of Arizona; Anderson, Raymond - Ray

Submitted to: Agricultural Water Management
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/11/2020
Publication Date: 5/22/2020
Citation: French, A.N., Hunsaker, D.J., Sanchez, C.A., Saber, M., Gonzalez, J.R., Anderson, R.G. 2020. Satellite-based NDVI crop coefficients and evapotranspiration with Eddy Covariance Validation for multiple durum wheat fields in the US Southwest. Agricultural Water Management. 239. https://doi.org/10.1016/j.agwat.2020.106266.
DOI: https://doi.org/10.1016/j.agwat.2020.106266

Interpretive Summary: A key aspect for efficient irrigation management relies on timely information about crop water requirements. A practical and widely used method to estimate crop water requirements is known as the FAO56 method, which estimates the water use of a crop by multiplying a crop coefficient by reference water use value that is determined from weather station data. However, because the crop coefficient changes dramatically over the season, it is very difficult to estimate every day. This study conducted research to determine the actual daily crop coefficient and water use of wheat crops using remote sensing information from satellites. This remote sensing approach was shown to estimate the daily measured water use very well, particularly during periods when the need for irrigation was the greatest. These results will lead to future development of this approach for providing reliable guidance for efficient irrigation management.

Technical Abstract: A three-year study was conducted to assess the ability of satellite-based vegetation index (VI) images to track evapotranspiration over wheat. While the ability of using VIs, notably with the Normalized Difference Vegetation Index (NDVI), to track vegetation growth has been well established, the operational capability to accurately estimate the crop coefficient (Kc) and crop evapotranspiration (ETc) at farm-scale from spaceborne platforms has not been widely studied. The study evaluated wheat ET over 7 sites between 2016 and 2019 in Yuma and Maricopa, Arizona, USA estimated by using Sentinel 2 and Venus satellites to map NDVI time-series for entire wheat cropping seasons, December to June. The basal crop coefficient (Kcb) was modeled by the NDVI time-series and the daily FAO56 reference ETo was obtained by near-by weather network stations. Eddy covariance (EC) stations in each field observed ETc during the same seasonal periods, and applied irrigation amounts were logged. The experiment found that remote sensing of NDVI and modeled Kcb accurately estimated Kc and crop ET during mid-season through senescence in most cases. However, NDVI-based estimation performed less well during early season (<60 days after planting), when observed ETc was highly variable due to frequent rain and irrigation at low crop cover. Mid-season Kc values observed for the seven wheat fields were from 0.92 to 1.14, and end of season Kc values ranged from about 0.20 to 0.40, in close agreement to values reported elsewhere. Seasonal VI-based transpiration and ETc values ranged from 467 to 618'mm, closely agreeing with seasonal EC data, which ranged 499–684'mm. Using the Venus sensor, the study in Maricopa in 2019 revealed that when augmented by a background soil water balance model, water stressed wheat can be detected mid-season with NDVI. This capability is specifically due to the sensor’s ability to provide well-calibrated images every 2 days. Findings from this study will help farmers, irrigators, and water managers use and understand the capabilities of visible near infrared remote sensing to track ETc from space.