Applied Water Science (Jun 2022)

WaSim model for subsurface drainage design using soil hydraulic parameters estimated by pedotransfer functions

  • Mphatso Malota,
  • Joshua Mchenga,
  • Brighton Austin Chunga

DOI
https://doi.org/10.1007/s13201-022-01699-z
Journal volume & issue
Vol. 12, no. 7
pp. 1 – 11

Abstract

Read online

Abstract The agricultural drainage engineering community is steadily shifting the design of subsurface drainage systems from the experience-based design approach to the simulation-based design approach. As with any design problem, two challenges are faced; firstly, how to determine all the input data required by the simulation model, and secondly to, a priori, anticipate what the performance of the designed system will be. This study sought to evaluate the performance of the WaSim model to simulate fluctuating water table depths (WTD), and drainage discharges (DD) in KwaZulu-Natal Province, South Africa. Saturated hydraulic conductivity (K sat), which is an input to the WaSim model, was estimated by the Rosetta computer program, based on soil particle size distribution data, bulk density, and soil water retention characteristics at pressure heads of – 33 and – 1500 kPa. performance of the WaSim model was statistically assessed using the coefficient of determination (R 2), coefficient of residual mass (CRM), mean absolute error (MAE), mean percent error (MPE), and the nash–sutcliffe efficiency (NSE). during the validation period, the WaSim model predicted WTDs with R 2, CRM, MAE, MPE, and NSE of 0.86, 0.003, 4.9 cm, 6.0%, and 0.98, respectively. In the same validation period, the model predicted DDs with R 2, CRM, MAE, MPE, and NSE of 0.57, 0.002, 0.30 mm day−1,11%, and 0.76, respectively. These results suggest that the use of Rosetta-estimated K sat data as inputs to the WaSim model compromised its accuracy and applicability as a subsurface drainage design tool. Owing to the relatively low R 2 value of 0.57, and that the WaSim model was empirically developed, we recommend further improvement on the calibration of the model for it to be suitable for application under the prevailing conditions. Also, in the absence of other means of determining K sat, we caution the use of Rosetta-estimated K sat data as inputs to the WaSim model for the design and analysis of subsurface drainage systems in KwaZulu-Natal Province, South Africa.

Keywords