H026-03
Validation of the SAFER ET model using Landsat 8 and Sentinel-2 images over commercial vineyards in California

Monday, 7 December 2020: 20:38
Virtual
Anderson Safre1, Ayman Nassar2, Mahyar Aboutalebi2, Alfonso F Torres-Rua2, John H Prueger3, Lynn McKee4, Joseph G Alfieri4, Lawrence Hipps5, Hector Nieto6, William Alexander White7, Maria Mar Alsina8, Luis Sanchez8, William P Kustas4 and Nick Dokoozlian8, (1)UNESP Sao Paulo State University, Irrigação e Drenagem, Botucatu, Brazil, (2)Utah State University, Department of Civil and Environmental Engineering, Logan, UT, United States, (3)U. S. Department of Agriculture, Agricultural Research Service, National Laboratory for Agriculture and Environment, Ames, IA, United States, (4)U. S. Department of Agriculture, Agricultural Research Service, Hydrology and Remote Sensing Laboratory, Beltsville, MD, United States, (5)Utah State University, Plants, Soils and Climate Department, Logan, UT, United States, (6)Complutum Tecnologías de la Información Geográfica S.L. (COMPLUTIG), Alcalá de Henares (Madrid), Spain, (7)USDA-ARS Hydrology and Remote Sensing Laboratory, Beltsville, MD, United States, (8)E&J Gallo, Modesto, United States
Recorded Presentation
Abstract:
The adequate estimation of evapotranspiration (ET) is critical for water management in commercial crops, especially in arid regions. Vineyards are one of the most challenging crops for ET estimation due to the interrow (soil/vegetation) effect, for example. Also, irrigation scheduling must be precise to induce water stress in order to reach the expected yield quality. Over the years, several models have been created to derive ET using meteorological data and reflectance on the different wavelengths obtained by satellite images. An ET model called SAFER (Simple Algorithm for Evapotranspiration Retrieviement) was developed in 2010 for the estimation of ET and energy balance using remote sensing images through the relation between NDVI, surface albedo, and surface temperature. This model has been applied in different locations and ecosystems in Brazil, showing excellent results when compared with eddy covariance measurements. The advantage of this model is that it requires few inputs from agrometeorological stations (Global radiation, air temperature, and ET0) and has two versions: one using thermal and other using only optical and near-infrared information. In this research, we evaluate the SAFER model for daily ET and energy fluxes estimation using Landsat 8 (30 m) and Sentinel 2 (10 m) images with EC flux measurements in vineyards located in Lodi, California, part of the USDA-ARL HRSL GRAPEX (Grape Remote sensing Atmospheric Profile & Evapotranspiration eXperiment) Project. The period analyzed was from 2013 to 2018, and a dataset of 125 images was included (91 Landsat 8; 34 Sentinel 2). Two flux towers located in different vineyards were used to assess the seasonal change in ET and energy fluxes. Monthly and annual statistic performance between EC and SAFER ET results is performed. A close examination is performed on months when the interrow crop was dominant (early season), and vines were dominant (late season) as well as midseason conditions. Additional analysis related to fractional ET (crop coefficient ET/ET0) derived by the EC towers and comparison with FAO56 is also included. The validation procedure and statistical performance of the SAFER ET model for both satellites are presented and discussed.
<< Previous Abstract | Next Abstract >>