An accurate estimation of the Evapotranspiration (ET) including spatial and temporal variations is very important in the agricultural sector for day-to-day operation, for planning of irrigated plots and for both short and long term water management. At local level, ET estimations can be obtained with high accuracy considering a soil-plant-atmospheric system. However, they cannot be directly extrapolated to regional scales mainly because different climatic and ground conditions may exist within the region analysed. The difficulty of making a spatial analysis of ET together with the lack of data can be minimised by using remote sensing techniques and data as an alternative to estimate both variables and parameters required to compute ET. The Surface Energy Balance Algorithm for Land (SEBAL) remote sensing procedure enables the estimation of ET_a and it has been successfully used in different environments.
In this research, data for the Flumen District and adjacent experimental plots, located in the Central Ebro Valley in Spain, was used. The SEBAL procedure is combined with meteorological-ground data, and validated with historical meteorological and ancillary ground data. As SEBAL was applied to a real agricultural area over four crop cycles, this study highlights the applicability of the procedure and its use under real conditions, which quite frequently are characterised by a lack of or inappropriate data, poor ground-truth data, local weather variations, field size, amongst others.
The validation at local level was made using accurate ET_o measurements from
lysimeters, as well as accurate ET_o computations from meteorological data using the Penman-Monteith method (FA0-56). However, at regional scale, ET_o measurements are not available, or if they are, the number of ET_o observations are highly limited. In this case, the limited availability of data necessitates the use of other ET_o methods such as the Hargreaves or Blaney-Criddle methods to evaluate the reliability of the ET_aSEBAL values on a regional scale.
Results showed that SEBAL, used together with ground-meteorological data, provides reliable ET_a estimations for sites in and adjacent to the Flumen District. ET_aSEBAL estimates on a pixel-by-pixel basis over a wide area are used to provide spatial coverage. Also, SEBAL demonstrates a high sensitivity to crop development and management practices. Thus, ET_aSEBAL allows the examination of water stress, making it possible to implement actions to improve the crop management as well as the irrigation schedule, nutrient management, pest control and other factors that help to increase yields. In addition, the ET_aSEBAL can be used as input data for different methods making SEBAL an important tool to achieve good results, for example in the present project computing irrigation volumes. These computed irrigation volumes closely matched with the actual irrigation volumes in the area.
Use of SEBAL as an alternative method of estimating ET_a, provided reliable data, and also using the results it was also possible to identify limitations and problems in the soil-plant-atmospheric system that affect the performance of SEBAL. This helps to establish a better control of the variables involved and to establish the best conditions for using SEBAL optimising the ET_aSEBAL results. The haze and Bochorno conditions were found in practice to be main factors that affect the ET_a, whereas the use of the SEBAL algorithm was found to be limited by the presence of wet and flooded soils. Under suitable conditions the SEBAL technique offers the opportunity to define spatial ET variability and the use of this for water allocation and management in real situations.