Resumen de Optimal irrigation scheduling combining water content sensors and remote sensing data
By 2025, the Food and Agriculture Organization of the United Nations predicts that two-thirds of the world population will experience water stress conditions. In addition, it is expected that world population will increase significantly during the following years. Agriculture is the largest consumer of fresh water, estimated 75%. The optimization of its use is therefore crucial for future. The main objective of irrigation scheduling is to guarantee maximum crop yield while saving significant amounts of water. The main difficulty to determine a correct and optimal irrigation scheduling strategy is due to complexity and variability of the Soil-Plant-Atmosphere continuum. In each field site, this continuous SPA has different patterns as a consequence of environmental conditions, crop, soil hydraulic properties and soil tillage. Several tools and methodologies are employed nowadays to monitor an improve irrigation scheduling. However, they all have advantages and limitations, or, each one is used without considering the information provided by the others, simplifying the system and avoiding information.
This thesis aims to improve irrigation scheduling by combining different tools available at the present time. This combination aims to answer different kind of questions when irrigation scheduling in a field site must be defined. For this, each tool has been used and supported by others. In addition, this thesis will discuss when and how these tools might be employed.
Firstly, we compare different tools and methodologies used to measure water content in an agriculture field. Specifically, we compared gravimetric measurements and water content sensors measurements with DISPATCH algorithm data. This algorithm is one of several algorithms who estimates soil surface water content using remote sensing data. The main goal was determining if remote sensing data can improve water content data measured by sensors. We found that DISPATCH algorithm is no capable to improve water content sensors measurements.
Secondly, we present a methodology where a simulation-optimization problem is solved in order to find an optimal irrigation scheduling strategy. This strategy might guarantee the maximum economic net margin. This methodology has been applied in a real field. The optimal irrigation scheduling of this field is compared with the traditional irrigation scheduling method, based water requirements. Results show that even though the traditional method supplies the volume of water evapotranspired, this methodology of scheduling irrigation is not enough to avoid crop water stress, compromising final crop yield. In this case, optimal irrigation strategy improves the final net margin in comparison the traditional method. We demonstrate that depending on soil properties, optimal irrigation scheduling is different.
Thirdly, we improved the irrigation scheduling in a field site where irrigation was applied with the same criteria. In this part of the thesis we employed NDVI remote sensing data, water content sensors and simulation models to determine the optimal irrigation. The improvement is based on management zones delineation with NDVI data. In this case, water content sensors are used to define if each management zone presents different water content patterns and to verify that when the field is divided into different management zones there is a gain in terms of water content uniformity. Finally, an optimal irrigation scheduling calendar is proposed to allow the consultants to take within-season irrigation decisions. Results show that management zones are dynamic during the growing season and the optimal irrigation scheduling might also be dynamic. In addition, we found that the uniform irrigation applied in the entire field, without considering the possible differences in soil properties, produced waterlogging in two management zones, therefore, transpiration decreases in comparison to the others.
