Comparison of irrigation automation algorithms for drip-irrigated apple trees

• Published in: Computers and electronics in agriculture Vol. 128; pp. 87 - 99
• Main Authors: Osroosh, Yasin, Peters, Robert Troy, Campbell, Colin S., Zhang, Qin
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.10.2016
• Subjects: Algorithms; Apples; Automatic irrigation scheduling; Drip irrigation; Irrigation; Plant-based approach; Seasons; Sensors; Soil dynamics; Thermal sensing; Thresholds; Trees; Wireless sensor network; Wireless sensor network; Automatic irrigation scheduling; Thermal sensing; Drip irrigation; Plant-based approach
• ISSN: 0168-1699; 1872-7107
• DOI: 10.1016/j.compag.2016.08.013

•We developed a decision support system integrated with a wireless sensor network.•We developed and field-tested seven irrigation scheduling algorithms for drip-irrigated apples.•We used the time-temperature threshold (TTT) method in apple trees for the first time.•The thermal-based treatments competed well with neutron probe (NP).•Thermal/weather-based irrigation scheduling reduced applied water by about 70% compared with conventional irrigation. Seven irrigation scheduling algorithms and an automatic control system along with a wireless network of soil, thermal and weather sensors were developed and assessed in Prosser, WA in the growing season of 2013. The system was comprised of six wireless sensor and valve actuating nodes installed across an apple orchard, a central base station made up of a transceiver connected to a laptop, and a graphical user interface (GUI). The irrigation algorithms/treatments included the time-temperature threshold (TTT), crop water stress index with dynamic threshold (CWSI), soil-based using granular matrix sensors (SOIL), weather-based using a temperature-only-based evapotranspiration (ET) model and soil water balance (WB), a combination of SOIL and WB (SL+WB), a conventional irrigation practice used in the region (CNTRL), and soil-based using a neutron probe (NP) as benchmark. Different treatments were compared based on the total irrigation water (It) applied during the season. They were also compared based on simplicity and expense for a grower to implement. Soil water content (θs) and stem water potential (ψstem) were monitored in a number of treatment plots. The total applied water for CNTRL was significantly higher than all other treatments (p<0.001). The thermal-based TTT and CWSI treatments applied the same amount of water as NP and WB (p<0.001). CWSI and TTT substantially reduced water applied (70%) while maintaining ψstem within the non-stressed range. In addition, θs in the treatment plots of TTT and CWSI did not exceed the maximum allowed deficit recommended for apple trees (MAD of 50%) showing a strong agreement with NP. The SOIL and SL+WB treatments resulted in tangible under-irrigation as leaf drop, decreased leaf turgidity, growth reduction and abnormally small fruits were seen. Among all the strategies, WB seemed to bear the characteristics of being economical, easy to implement and fairly accurate. Our preliminary results also support the use of wireless sensor network for automatic irrigation management of drip-irrigated apple trees.