Real-time optimization of irrigation of citrus fields near Valencia (Spain)
Irrigated agriculture is very important for supporting the huge population of mankind. Considering the limited global water resources, irrigation scheduling is needed to reduce water use while maintaining the agricultural production. In this study we were concerned with the real-time automatic control of irrigation, which calculates daily water allocation by combining information from soil moisture sensors and a land surface model. The main novel contribution of this work was the incorporation of soil moisture measurements to improve the estimate of the soil water status, and the use of a sophisticated land surface model for irrigation scheduling. Although near real-time irrigation scheduling is commonly applied in practice, it is mostly based on using information from soil moisture sensors, without hydrological or land surface model, and without data assimilation.
We tested near real-time scheduling of irrigation for a site with citrus trees near Picassent, Valencia. The irrigation scheduling was carried out operationally for six fields, and other strategies for irrigation scheduling were applied to other fields (FAO-based scheduling for three fields and scheduling by farmer´s experience for two fields). The estimates of soil and plant status at the site were made by combining soil moisture measurements and predictions by the Community Land Model (CLM) using sequential data assimilation (DA). The LETKF (Local Ensemble Transform Kalman Filter) was chosen to assimilate soil water content measured by FDR (Frequency Domain Reflectometry) into CLM and improve the initial (soil moisture) conditions for the model run. In addition, predictions by the GFS (Global Forecast System) model were used as atmospheric input data for CLM to predict short-term soil moisture contents. The difference between predicted and target soil water content was defined as the water deficit, and the irrigation amount was calculated by the integrated water deficit over the root zone. The corresponding irrigation time to apply the required water was introduced in SCADA (supervisory control and data acquisition system) for each citrus field.
During the real-time irrigation campaigns in Valencia in 2015 and 2016, the applied irrigation amount, stem water potential and soil moisture content were recorded. The data indicated that averaged over both years 18.5% less irrigation water was needed for the CLM-DA scheduled fields than for the traditionally irrigated fields. Stem water potential data and soil moisture recordings indicated that the CLM-DA fields were not suffering from water stress during the irrigation period. Even though the CLM-DA fields received less irrigation water, the orange production was not significantly suppressed either. Overall, our results show that the CLM-DA method is attractive given its automated approach, ease of incorporation of on-line measurements and ensemble based predictions of soil moisture evolution. In addition, it showed water saving potential compared to other traditional irrigation methods.
Contact person:
Dazhi Li
Agrosphere (IBG-3)
Forschungszentrum Jülich GmbH
Leo Brandtstrasse
52425 Jülich
Germany
Tel. +49 2461 61 96647
E-mail: d.li@fz-juelich.de
Relevant publications:
Han, X., H.J. Hendricks Franssen, M.A. Jimenez-Bello, R. Rosolem, H. Bogena, F. Martinez-Alzamora, A. Chanzy, and H. Vereecken. 2016. Simultaneous soil moisture and soil properties estimation for a drip irrigated field by assimilating cosmic ray neutron intensity. Journal of Hydrology 539, 611-624.
Han, X., H.J. Hendricks Franssen, R. Rosolem, R. Jin, X. Li, and H. Vereecken. 2015. Correction of systematic model forcing bias of CLM using assimilation of cosmic-ray neutrons and land surface temperature: a study in the Heihe catchment, China. Hydrology and Earth System Sciences 19(1), 615-629.