Improving field-scale crop actual evapotranspiration monitoring with Sentinel-3, Sentinel-2, and Landsat data fusion

• DOI: 10.1016/j.jag.2023.103587
• Publication Date: 2023-11-29T21:29:54Z
• Authors: Radoslaw Guzinski; Héctor Nieto; Rubén Ramo Sánchez; Juan Manuel Sánchez; Ihab Jomaa; Rim Zitouna-Chebbi; Olivier Roupsard; Ramón López-Urrea
• Keywords: Earth observation

One of the primary applications satellite Land Surface Temperature (LST) observations lies in their utilization for modeling actual evapotranspiration (ET) agricultural crops, with goals monitoring and enhancing irrigation practices improving crop water use productivity, as stipulated by Sustainable Development Goal (SDG) indicator 6.4.1. Evapotranspiration is a complex dynamic process, both temporally spatially, necessitating LST high spatio-temporal resolution. Presently, none existing spaceborne thermal sensors can provide quasi-daily field-scale observations, prompting development methods data fusion (thermal sharpening) from various shortwave to meet this requirement. Previous research has demonstrated effectiveness combining shortwave-multispectral Sentinel-2 thermal-infrared Sentinel-3 derive daily, ET estimates. However, these studies also highlighted limitations capturing distinct contrast between cooler irrigated areas hotter, adjacent dry regions. In study, we aim address limitation incorporating information on spatial variability observed Landsat satellites into without being constrained infrequent or cloudy while retaining longwave radiance emission captured sensor at its native Two approaches are evaluated, individually complementary combination, validated against situ measurements. The best performing approach, which leads reduction root mean square error up 1.5 K when compared previous research, subsequently used estimate parcel-level evapotranspiration. process undergone improvements regarding gap-filling input output data, datasets code implementation. resulting using lysimeters eddy covariance towers Spain, Lebanon, Tunisia, Senegal minimal overall bias (systematic underestimation less than 0.07 mm/day) low (down 0.84 fully global datasets. enhanced sharpening methodology agnostic should remain relevant upcoming missions accuracy modeled fluxes encouraging further Sentinel satellites, other Copernicus SDG