Estimatimation of crop water requirements by remote sensing for variable rate applications. operational case in a central pivot of wheat

In Mediterranean areas, where water scarcity is the main limiting factor, applying good practices in the use of water for irrigation is crucial in order to maximize benefits for farmers and protect the resource. Furthermore, energy costs derived from water pumping from groundwater is one of the most important expenses for farmers in our study area, the South-East of Spain. Variable Rate Irrigation is a promising technique to apply the required water, but VRI faces the challenge to know accurately the crop water requirement distribution in space and time.

The objective of this work is twofold: Firstly, to demonstrate through a practical case the optimization of the irrigation water in an operativity level managing the variability of the plot using time series of free satellite images currently in orbit. Secondly, to put into practice the technology (SicoP system) developed by ACOEMAN that allows the pivot to apply variable rate at medium cost for farmers.

The case study was carried out in a commercial wheat plot of 60ha, irrigated by a central pivot endowed with the SicoP technology, during the campaign of 2018-2019.  The SicoP pivot technology allows to implement a variable angular speed for each sector. The pivot circle was divided into 36 sectors of 10 degrees each. Every Thursday during the growing cycle the crop water requirements were estimated per sector by means of remote sensing and meteorological data by the decision support system developed by AgriSat Iberia as consultant company. Thus, the system applied the irrigation water requirement per sector, calculated through a simplified soil water balance.

The estimation of the actual crop water requirements spatially distributed at 30x30 meter (3x3 pixel) resolution has been based on NDVI-Kc forecasting methodology. The high temporal and spatial resolution provided by free images from satellites Sentinel 2A and Sentinel 2B combined with Landsat 8 images allows the implementation of a remote sensing-based operational approach for this variable rate decision support system.

This paper includes a comparative analysis of the differences between the water volume applied by homogeneous rate, 1 per plot and week, and the variable rate irrigation, 36 rates per plot and week, using the same EO-based methodology. A yield map was obtained by using a yield-monitoring device implemented into the combine harvester.

First promising results regarding the optimization of the use of water have been demonstrated going from 1 irrigation decision in 60ha per week, to 36 irrigation decision per week, one per 1.6ha sector. Modest savings in water volumes at the end of the growing cycle have been observed. This map shows no additional increase of yield spatial variability due to the use of VIR.  Some problems were encountered when the climate conditions were not appropriate for irrigation, mainly high wind speed. The system has reached a high operativity level ready for adoption by farmers.