1. Introduction
The continuous increase in water demand is subject of concern (Cosgrove & Loucks, 2015; Boretti & Rosa, 2019) and demands efficient management systems, particularly in semiarid regions (Sun et al., 2019), which account for 15% of the global land area and feed 14% of the global population. Besides, semiarid ecosystems are very fragile and sensitive to climate change (Safriel & Adeel, 2005).
The Brazilian semiarid zone is characterized by an extremely negative atmospheric water balance (annual precipitation of 549 mm against potential evaporation of 2,500 mm) and a considerable spatial and temporal variability in water supply. This area presents only two annual seasons: dry and rainy (de Araújo & Piedra, 2009; de Figueiredo, de Araújo, Medeiros, & Costa, 2016; Pinheiro, Metselaar, Lier, & Araújo, 2016). These features contrast with other semiarid areas, such as the Spanish Northern Plateau, where precipitation distribution is relatively uniform throughout the year and presents well-defined seasons (Calama et al., 2019). These differences may lead to a distinct spatial and temporal behaviour of the actual evapotranspiration (ETa) in those semiarid areas.
Evapotranspiration (ET) is one of the most challenging components of the water balance equation to be measured (Castelli et al., 2018). It can be quantified using lysimeters or aerodynamic methods (e.g., Bowen ratio and Eddy Covariance), but such methods only represent the specific area around the experimental setup. The spatial variation of evapotranspiration on a large scale can be better captured by models based on remote sensing (Jaafar & Ahmad, 2020). The high spatial resolution and long registration period of a USGS Landsat permits to assess a spatiotemporal ET series of vegetated and non-vegetated surfaces (Chen & Liu, 2020; Jaafar & Ahmad, 2020).
Thermal-energy balance models that use remote sensing, such as SEBAL (Surface Energy Balance Algorithm for Land), have been successfully applied to assess ETa at field scale in many areas of the world. Teixeira, Bastiaanssen, Ahmad, and Bos (2009) applied SEBAL to estimate evapotranspiration in irrigated agriculture and Caatinga vegetation in the semiarid region of the São Francisco River basin, north-eastern Brazil; Losgedaragh and Rahimzadegan (2018) used it to compute evaporation from the freshwater bodies; the same authors (2018), as well as Rahimzadegan and Janani (2019), applied SEBAL to agricultural lands in a semiarid climate in Iran; Senkondo, Munishi, Tumbo, Nobert, and Lyon (2019) used it in a sub-tropical region in Tanzania’s Kilombero Valley; whereas Gobbo et al. (2019) and Grosso et al. (2018) applied it to assess ET at irrigated and non-irrigated maize fields in Italy.
In the years to come, a significant reduction of precipitation and a temperature increase can be expected over global typical semiarid regions as a result of climate change (Yang, Zhang, Hao, & Yue, 2019). This trend may influence both the atmospheric and vegetation water demand, that is, the ETa. Therefore, it is crucial and challenging to research for a better hydrological understanding of water-balance changes that may occur in semiarid ecosystems, such as Caatinga (Brazil) and Tierra de Pinares (Spain). The aim of this study is, hence, to compare spatial and temporal patterns of actual evapotranspiration, as well as to identify temporal evapotranspiration trends in both semiarid forests: Caatinga (Brazil) and Tierra de Pinares (Spain).