Evaluation of Infiltration Models in Soils of Two Land Use Systems in Calabar, Nigeria

Abstract

Recently, the importance of the infiltration process in agriculture and the environ- ment has resulted in an upsurge of interest by soil and water scientists to model the process for quantitative application. A study was conducted on the University of Cal- abar Teaching and Research Farm, Calabar to evaluate the effect of oil palm (OP) and arable farm (AF) land use systems on the Green-Ampt (GA), Philip (P), Kostiakov (K), Horton (H) and Mezencev (MZ) infiltration models, as well as the applicability or efficiency of the models to predict infiltration into the soils. Infiltration data were obtained with double-ring infiltrometer, and the parameters of the models were obtained through curve-fitting. Model accuracy was evaluated with the Willmott’s index of agreement (W), chi-square (X2), coefficient of determination (R2), root mean square error (RMSE) and mean error (ME) test statistics. The results showed that soil under oil palm had measured cumulative infiltration and infiltration rate of 72.81 cm and 14.10cm/hr while arable farm had 74.76 cm and 12.92 cm/hr, respectively. The cumulative infiltration predicted by Philip and Kostiakov models were very close to the field data for OP and AF. Horton and Mezencev models underestimated the infiltration process because their ME values were negative while Green–Ampt, Kostia- kov, and Philip overestimated the infiltration process as they had positive ME values. In terms of accuracy and applicability, the order of performance was P>K>MZ>GA>H. Therefore, the Philip and Kostiakov models could be used to pre- dict infiltration into the soils, but that the Philip model was superior to the Kostiakov model for the University of Calabar Teaching and Research Farm and similar soils in other ecologies.