Calibration of Existing Infiltration Models on Different Amended Soils

摘要

Organic amendments can affect soil properties. The predictive power of selected waterinfiltration models in organic amended soils were evaluated. Ten (10) infiltration models consisting of five (5)empirical (Philip (PH), Kostiakov (KT), Modified Kostiakov (MK), Kostiakov-Lewis (KL) and Natural ResourceConservative Service (NRCS)), Three (3) physically based (Green-Ampt (GA), Smith-Parlange (SP), Talsma-Parlange (TP)) and two semi-empirical ( Swartzendruber (SW) and Horton (HT)), were evaluated for soilsamended with cow dung, poultry litter and Pig dung. Ability of the models to accurately prediction the measuredcumulative infiltration. The study was carried out at University of Uyo experimental plot at Uyo, Akwa Ibom.The field was ploughed to a depth of 20cm in order to mix the manure with the soil. Soil samples were takenfrom each strip, and the tests were repeated two weeks, four weeks and six weeks after manure application. Theamendment with Cow dung increased cumulative infiltration by 56.05%, Poultry litter increased cumulativeinfiltration by 36.43% and Pig dung manure increased cumulative infiltration by 2.6%. The statistical modelwas used to evaluate the model performance. The coefficient of determination ???? between the models simulatedcumulative infiltration and field measured cumulative infiltration ranged from 0.931 to 0.998. The value of themodelling efficiency (??) index ranged from -0.048 to 0.998 while the Mean Absolute Error (??????) rangesfrom -0.955 to 5.454. The values of the coefficient of residual mass (??????) ranged from -1.320 to 0.572. Sevenmodels (NRCS, Philips, Kostiakov, Kostiakov-Lewis, Modified Kostiakov, Talsma-Parlange andSwartzendruber’s model) under-predicted while three models (Green-Ampt, Smith-Parlange and Horton) overpredictedcumulative infiltration. The NRCS’s model had the overall best performance, Kostiakov model hadbest performance amongst the empirically based models, the Talsma-Parlange’s model had the bestperformance in the physically based models, and Horton’s model had the best performance in the semiempiricalgroup respectively. The NRCS’s model were found to be most appropriate for cow dung, Poultrylitter, and the Control while, Kostiakov model were found to be most suitable in the Pig dung amended soil.