Inverse modelling in estimating soil hydraulic functions: a Genetic Algorithm approach

摘要

The practical application of simulationmodels in the field is sometimes hindered by the difficulty of deriving the soilhydraulic properties of the study area. The procedure so-called inversemodelling has been investigated in many studies to address the problem wheremost of the studies were limited to hypothetical soil profile and soil coresamples in the laboratory. Often, the numerical approach called forward-backwardsimulation is employed to generate synthetic data then added with random errorsto mimic the real-world condition. Inverse modelling is used to backtrack theexpected values of the parameters. This study explored the potential of aGenetic Algorithm (GA) to estimate inversely the soil hydraulic functions in theunsaturated zone. Lysimeter data from a wheat experiment in India were used inthe analysis. Two cases were considered: (1) a numerical case where theforward-backward approach was employed and (2) the experimental case where thereal data from the lysimeter experiment were used. Concurrently, the use of soilwater, evapotranspiration (ET) and the combination of both were investigated ascriteria in the inverse modelling. Results showed that using soil water as acriterion provides more accurate parameter estimates than using ET. However,from a practical point of view, ET is more attractive as it can be obtained withreasonable accuracy on a regional scale from remote sensing observations. Theexperimental study proved that the forward-backward approach does not take intoaccount the effects of model errors. The formulation of the problem is found tobe critical for a successful parameter estimation. The sensitivity of parametersto the objective function and their zone of influence in the soil column aremajor determinants in the solution. Generally, their effects sometimes lead tonon-uniqueness in the solution but to some extent are partly handled by GA.Overall, it was concluded that the GA approach is promising to the inverseproblem in the unsaturated zone. style="line-height: 20px;">Keywords. Genetic Algorithm, inverse modelling, Mualem-Van Genuchten parameters, unsaturated zone, evapotranspiration, soil water