A Linking Test that establishes if groundwater recharge can be determined by optimising vegetation parameters against soil moisture

The impact of afforestation/deforestation on groundwater recharge can be predicted by using one-dimensional soil-vegetation water flow models based on Richards' equation. However simulations depend upon parameters that are not easily measurable. Pollacco et al. (2008) showed that the hydraulic parameters can be determined, if the vegetation parameters are known, by fitting simulated time series of soil moisture profiles to those measured in situ. This paper presents a case study to determine if the interception and crop factor parameters can tentatively be calibrated by fitting soil moisture profiles. Synthetic data were used and the other vegetation parameters and the soil hydraulic parameters were assumed to be known. We applied and improved the Linking Test developed by Pollacco et al. (2008) to look for links between the parameters that need to be calibrated, and thus to investigate whether inverse modelling is feasible, which depends on the accuracy of the calibration data. The Linking Test established that interception and evapotranspiration parameters are linked and, therefore, uncertainty on interception compensates for uncertainty on evapotranspiration. Thus in spite of a good match between observed and simulated soil moisture data, inverse modelling is unfeasible. This is true even if the interception or the crop factor parameters are known, because an error on interception or evapotranspiration will be compensated by an error on groundwater recharge without affecting soil moisture. This paper recommends that vegetation parameters should not be calibrated by optimisation against soil moisture data.