Testing the Goodness of Fit of Selected Infiltration Models on Soils with Different Land Use Histories

Aug 29, 2026

Six infiltration models, derived by reformulating parameters of the classical Kostiakov (1932) and Philip (1957) equations, were evaluated for describing water infiltration into highly permeable sandy soils in the Nsukka plains of SE Nigeria. Infiltration measurements from double-ring infiltrometers were collected on Kandic Paleustult soils under different land-use histories and maize-yield treatments. Cumulative infiltration ranked highest on tilled-mulched plots and lowest on continuous pasture, and fitted parameters varied significantly. Model performance was assessed via R² and deviations between experimental and predicted cumulative infiltration, proposing Modified Kostiakov or Modified Philip for routine characterization on high-infiltration soils.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages10

Reading Time25 min

AccessFree to download

Abstract

Reference

Introduction

Materials and methods

Site history and land use

# TESTING THE GOODNESS OF FIT OF SELECTED INFILTRATIONMODELS ON SOILS WITH DIFFERENT LAND USE HISTORIES J.S.C.Mbagwu¹International Centre for Theoretical Physics,Trieste,Italy. # ABSTRACT Six infiltration models,some obtained by reformulating the fitting parameters of theclassical Kostiakov(1932)and Philip(1957)equations,were investigated for their abilityto describe water infiltration into highly permeable sandy soils from the Nsukka plainsof SE Nigeria.The models were Kostiakov,Modified Kostiakov(A),Modified Kosti-akov(B),Philip,Modified Philip(A)and Modified Philip(B).Infiltration data wereobtained from double ring infiltrorneters on field plots established on a Kandic Paleustult(Nkpologu series)to investigale the effects of land use on soil properties and maize yieldThe treatments were;(i)tilled-mulched(TM),(ii)tilled-unmulched(TU),(iii)untilled-mulched(UM),(iv)untilled-unmulched (UU)and(v)continuous pasture(CP).Cunula-tive infiltration was highest on the TM and lowest on the CP plots,All estimated modelparameters obtained by the best fit of measured data differed significantly among thetreatments.Based on the magnitude of R²values,the Kostiakov,Modified Kostiakov(A),Philip and Modified Philip(A)models provided best predictions of cumulative infil-tralion as a function of time.Comparing experimental with model-predicted cumulativeinfiltration showed,however,that on all treatments the values predicted by the classicalKostiakov,Philip and Modified Philip(A)models deviated most from experimental dataThe other models produced values that agreed very well with measured data.Consider-ing the ease of determining the fitting parameters it is proposed that on soils with highinfiltration rates,either Modified Kostiakov model(1=Kt+Ict)or Modifed Philipmodel(I=St¹/²+Ict),(where I is cumulative infiltration,K,the time coefficient,t,time elapsed,a'the time exponent,Ie the equilibrium infiltration rate and S,the soilwater sorptivity},be used for routine characterization of the infiltration proces. MIRAMARE-TRIESTE October 1993 # REFERENCE 1 Introduction Infiltration of water into soils is one of the most studied of all the important hydrologicprocesses occurring at the soil-atmosphere interphase.It is a dynamic property whichchanges drastically both temporally and spatially in response to changes in soil and cropmanagement practices.Gumbs and Warketin(1972)reported large variatjons in soilinfiltration capacity following small changes in bulk density,and Mbagwu(1987,1990)and Lal et al.(1980)observed much higher curnulative infiltration,infiltration rate andtime to attain equilibrium infiltration rate on mulched than unmulched plots irrespectiveof the tillage treatments. Many attemps have been made to mode the infiltration process from physically-and/or empirically-derived equations.Among these,two models that have been usedmore frequently because of their simplicity and ease of computing their fitting parametersare those of Kostiakov(1932)and Philip(1957),(see Table 1).Davidoff and Selim(1986)examined the validity of these two models for describing the infiltration processon soils with varying cropping histories in Louisiana,USA,whereas Mwendera and Feyen(1993)examined their applicability to freshly tilled surfaces with various tillage-inducedmicro-structure.Philip(1969)and Dunin(1976)observed that for theoretically longtime intervals the transmissivity term(A)in the Philip model approximates the soilfeld-measured saturated hydraulic conductivity(Ks).Other workers(Gish and Starr,1983;Maller and Sharma 1984;Swartzendruber and Youngs,1974)have also assertedthat A=Ks will not introduce serious errors in estimating cumulative infiltration as afunction of time.Collis-George(1977),Skaggs and Khaleel(1982),Smiles and Knight(1976)and Whisler and Bouwer(1970)showed,however,that assuming A=Ks over-predicted cumulative infiltration for large times.Furthermore,Jaynes and Giford(1981),Skaggs et al;(1969)and Watson(1959)who have examined

FAQs

Which infiltration models were tested in this study?

The study tested six models: Kostiakov, Modified Kostiakov (A and B), Philip, Modified Philip (A and B). Several were obtained by reformulating fitting parameters from the classical equations.

How were the infiltration measurements obtained and what soil conditions were used?

Water infiltration was measured using double-ring infiltrometers on field plots with Kandic Paleustult (Nkpologu series) soils at Nsukka in SE Nigeria.

How did land use histories affect cumulative infiltration and model fitting?

Cumulative infiltration was highest under tilled-mulched treatment and lowest under continuous pasture. Estimated model parameters from the best fits differed significantly among treatments, and some models deviated more from measured values than others.

Aug 29, 2026

Uploaded by Evangeline