1. Introduction
Agriculture consumes roughly seventy percent of the freshwater withdrawn worldwide, and in semi-arid countries such as Jordan the share is even higher while renewable supplies per person are among the lowest in the world. Most smallholder farmers decide when to irrigate from habit, the appearance of the crop or a fixed calendar. Direct measurement of soil water would allow irrigation to follow crop demand (Allen et al., 1998), but commercial soil moisture systems remain too expensive for farms of one to three hectares.
In this study we ask whether a capacitive probe built from inexpensive components can be accurate enough to schedule irrigation, and whether doing so saves water without harming yield.
2. Materials and Methods
2.1 Study site and design
The experiment ran from March to August 2025 on 18 plots of drip-irrigated tomato (Solanum lycopersicum) in the Jordan Valley. Table 1 summarises the soils. Nine plots were irrigated according to probe readings and nine followed the farmers’ usual calendar.
| Soil class | Sand (%) | Silt (%) | Clay (%) | Plots |
|---|---|---|---|---|
| Sandy loam | 62 | 24 | 14 | 6 |
| Loam | 41 | 38 | 21 | 7 |
| Clay loam | 33 | 31 | 36 | 5 |
2.2 Sensor and calibration
The probe measures the capacitance of the soil between two electrodes and converts it to volumetric water content. Each probe was calibrated in the field against a time-domain reflectometry (TDR) reference (Topp et al., 1980) at an air-dry and a field-capacity point.
3. Results
After calibration the probe agreed closely with the reference (RMSE = 2.1% volumetric water content, R² = 0.94, n = 1,260 paired readings). Plots irrigated by probe received 27% less water over the season than calendar plots, and marketable yield did not differ significantly (p = 0.31).
| Treatment | Water applied (mm) | Yield (t/ha) |
|---|---|---|
| Calendar (farmer practice) | 612 | 78.4 |
| Probe-based scheduling | 447 | 77.1 |
4. Discussion
The saving observed here is in line with reviews of sensor-based irrigation (Robinson et al., 2008). Sensor drift appeared late in the season in the two sandiest plots, which suggests re-calibrating once per season. The study covers one crop in one season, so the results should be confirmed on other crops and soils.
5. Conclusion
A probe costing under US$12 can support irrigation scheduling with accuracy sufficient for tomato production, saving about a quarter of the water without a measurable yield penalty. Wider adoption would require training and a simple calibration routine.
Acknowledgements
The authors thank the participating farming families in the Jordan Valley.