International Journal of Applied Earth Observations and Geoinformation 98 (2021) 102298 Available online 29 January 2021 0303-2434/© 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Remote sensing-based actual evapotranspiration assessment in a data-scarce area of Brazil: A case study of the Urucuia Aquifer System Bruno C´ esar Comini de Andrade a, * , Eber Jos´ e de Andrade Pinto b, c , Anderson Ruhoff a , Gabriel B. Senay d a Instituto de Pesquisas Hidr´ aulicas, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil b Escola de Engenharia, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil c Serviço Geol´ ogico do Brasil-CPRM, Belo Horizonte, Brazil d U.S. Geological Survey (USGS) Earth Resources Observation and Science (EROS) Center, North Central Climate Adaptation Science Center, Fort Collins, CO 80523, USA A R T I C L E INFO Keywords: SSEBop SMAP GRACE MODIS ABSTRACT The large groundwater reserves of the Urucuia Aquifer System (UAS) enabled agricultural development and economic growth in the western Bahia State, in northeastern Brazil. Over the last several years, concern has grown around the aquifer’s diminishing water levels, and water balance (WB) studies are in demand. Considering the lack of measured actual evapotranspiration (ET a ), a major component of the water cycle, this work uses the Operational Simplifed Surface Energy Balance (SSEBop) model to estimate ET a , and compares it to basin-scale estimates from the Soil Moisture Accounting Procedure (SMAP) monthly model and from an annual WB closure method, based on gridded meteorological data and the Gravity Recovery and Climate Experiment (GRACE) product. Additionally, a comparative assessment of different versions of the SSEBop parameterization was per- formed. Moderate Resolution Imaging Spectroradiometer (MODIS) imagery was used to implement eight different versions of the SSEBop algorithm over the UAS between 2000 and 2013. SSEBop and SMAP ET a yielded similar seasonal patterns, with correlation coeffcient (r) up to 0.65, mean difference (MD) of 0.8 mm/month and mean absolute difference (MAD) of 18.5 mm/month. Comparison of SSEBop annual ET a estimates to annual SMAP and WB closure estimates yielded low MD (12.1 and 7.3 mm/year, respectively) and MAD (82.5 and 82.8 mm/year, respectively), but also low r values (0.00 and 0.37, respectively). The comparison of the different SSEBop versions indicated the need to incorporate a calibration step of the aerodynamic heat resistance (r ah ) parameter. SSEBop results were also used for land cover and drought monitoring. Analysis indicates that agri- culture, associated with an increasing trend of atmospheric evaporative demand, is responsible for the decrease in groundwater levels and streamfow in the studied time period. 1. Introduction The Urucuia Aquifer System (UAS) is one of the largest groundwater reservoirs of Brazil, with an area close to 120,000 km 2 . Water from its sandstone formations supplies the S˜ ao Francisco River and is partially responsible for streamfow maintenance during the dry season (ANA, 2018). Over the last decades, the UAS native vegetation was gradually replaced by crop plantation and livestock farming. In the past few years, concern around the aquifer’s ability to replenish itself has grown, as a decrease in the local rivers’ basefow and terrestrial water storage was observed (Gonçalves et al., 2019; Pousa et al., 2019). Therefore, studies about the UAS water balance are needed, since monitoring of important water cycle processes, such as actual evapotranspiration (ET a ), is scarce. In a river basin, the main source of water loss is ET a , a process that comprises evaporation from soil, plants and water bodies; and transpi- ration from plants (Tateishi & Ahn, 1996; Fisher et al., 2017). These two sub-processes occur simultaneously, making it diffcult to distinguish one from the other (Allen et al., 1998). For this reason, ET a is often addressed as a single process. Knowledge of ET a rates is relevant both for estimates of crop consumption, at a local scale, and for assessment of water availability, at a regional scale. Land ET a is measured locally by fux towers and lysimeters and can * Corresponding author. E-mail address: cominideandrade@gmail.com (B.C.C. de Andrade). Contents lists available at ScienceDirect International Journal of Applied Earth Observations and Geoinformation journal homepage: www.elsevier.com/locate/jag https://doi.org/10.1016/j.jag.2021.102298 Received 23 October 2020; Received in revised form 30 December 2020; Accepted 31 December 2020