EBMAN-HP: A parallel model for simulation of sensor-based ebb-and-flow subirrigation systems

Ebb-and-flow irrigation system is a closed-loop efficient subirrigation system. In this study, a numerical model (EBMAN-HP) has been presented for simulation of all components (variations of water depth in supply tank and concrete floor/tank) and all phases of flood-floor/bench ebb-and-flow subirrigation systems. The model benefits from a fine-tuned computational algorithm for hysteresis module. The model can simulate both time-specified and sensor-based irrigation scheduling. Since ebb-and-flow irrigation system incorporates numerous pots, Richards’ equation should be solved for several pots to obtain sufficient understanding of the whole system. Therefore, the proposed model benefits from OpenMP parallel programming to speed up the execution time. Besides, a novel parallel TDMA solver have been presented that accelerates the computation speed by breaking a large system of equations into several simultaneously-solved portions. The model has been validated and verified against several analytical, numerical and experimental test cases. The results showed hysteresis module can completely remove artificial pumping error in two critical test cases. The parallel TDMA solver was shown to be able to reach to the speedup of about 90 %. The model was shown to perform faster than Hydrus-1D even in serial mode for coarser grids (about 52 % faster in average of 8 test cases) and similar to Hydrus-1D for dense grids (about 6 % faster in average of 4 test cases) with the perfect agreement (NSE between 0.999 and 1.000 and the average difference in MBE less than 0.1 % for 12 cases). Parallel model could boost the models’ performance to about 500 % using 6 processors. Finally, comprehensive illustrative example has been shown to present almost all capabilities of model.