An assessment of shallow water tables and the development of appropriate drainage design criteria for sugarcane in Pongola, South Africa.
South Africa, in common with all countries with arid or semi-arid climatic conditions, is facing the consequences of irrigation development without effective subsurface drainage. The quality of irrigation water is also decreasing and hence more water is required for leaching. This is resulting in low irrigation water productivity, as a consequence of shallow water tables, thus limiting crop growth. This study investigated the nature and causes of shallow water table problems in the sugarcane fields of Pongola, South Africa. The DRAINMOD model was also assessed for its reliability to be used as drainage design tool in the area. A water table map of a 32 ha sugarcane field was generated using groundwater table data monitored in 36 piezometers from September 2011 to February 2012. Nearly 12 % of the 32 ha sugarcane field was found to be affected by shallow water tables of less than the 1.0 m Design Water Table Depth (WTD). The inability of the adopted Drainage Design Criteria (DDC) to cope with drainage needs was found to be the cause of the poor drainage problem. On the other hand, analysis of WTDs in a field with a poorly-maintained subsurface drainage system confirmed that the drainage problem is exacerbated by poor drainage maintenance. It was recommended that the subsurface DDC in the area be revisited and that timely maintenance also be provided The DRAINMOD model was calibrated and verified using actual WTD and Drainage Discharge (DD) data. The model evaluation results revealed that the DRAINMOD model can reliably predict WTDs, with a Goodness of fit (R2), Mean Absolute Error (MAE) and Coefficient of Residual Mass (CRM) of 0.826, 5.341 cm and -0.015, respectively. Similarly, the model evaluation results in predicting DDs were also good, with R2, MAE and CRM of 0.801, 0.181 mm.day-1 and 0.0004, respectively. A further application of the validated model depicted that drain pipes installed at depths ranging from 1.4 m to 1.8 m and a spacing ranging from 55 to 70 m, with a design discharge of 2.5 to 4.2 mm.day-1, were adequate in ensuring safe WTDs between 1.0 and 1.5 m in clay-loam soil. On the other hand, drain depths ranging from 1.4 to 1.8 m and spacing between 25 and 40 m, were found to be appropriate in maintaining WTDs between 1.0 and 1.5 m in clay soil, with drainage design discharge ranging from 2.5 to 5.1 mm.day-1. These findings suggest that the current drain spacing needs to be reduced, in order to maintain the 1 m design water table depth. Finally, for the adoptability of the DRAINMOD model in the area, the Rosetta program, a component of the HYDRUS-2D, was tested for its reliability in estimating saturated hydraulic conductivities required by the DRAINMOD model. Results of the investigation revealed that the program can reliably be used to estimate saturated hydraulic conductivities from easily accessed soil data (% sand, silt, clay and soil bulk density), with R2, MAE and CRM of 0.95, 0.035 m.day-1 and -0.031, respectively. Nonetheless, calibration of the DRAINMOD model based on saturated hydraulic conductivity estimated by the Rosetta program was recommended. The findings of this research will form the basis for implementing an agricultural drainage policy that will ensure sustainable rain-fed and irrigation crop production systems in South Africa.