Modelling chemical and biological reactions during unsaturated flow through sandy soils

Predicting chemical and biological reactions during unsaturated flow through soils is complex, and the accuracy of model predictions is open to question. Objectives were to test how accurately the transport of nutrients (Ca2+, Mg2+, K+, NH4+, and NO3–) in soils can be predicted when differing amounts of nutrients or acid are added. Undisturbed columns of 2 sandy surface soils from arable sites in Darmstadt, Germany, and Sohar, Oman, were irrigated for 360 and 376 days with 3 mm/day at 10°C. Three treatments were carried out: the columns were irrigated with differing fertilising solutions (Fert-1 or Fert-2 treatments) or with HCl. The model PHREEQC2 was used to calculate 1-dimensional transport, inorganic complexation, dissolution or precipitation of CaCO3, multiple cation exchange, and nitrification. We compared 3 model variants: one with no adjustable parameters, a second with optimised nitrification rates and pCO2, and a third with an additional optimisation of the exchange coefficients. In model variant v2 and v3, the Fert-1 treatment was used for calibration. Model variant v1 was of little use in predicting cation transport in soils. The second was more appropriate for both soils, but an optimisation of the exchange coefficients (model variant v3) was required for a more accurate description (Fert-1) and prediction (Fert-2 and HCl) of ion transport. The included proton buffering reactions resulted in a moderately accurate prediction of pH and release of ions. Nitrification in the Fert-1 and Fert-2 experiments considered in the model variants accounted for a range of 0.03–0.04 µmol N/(day.g soil). Overall, our results indicate that calibration experiments are required for a reliable assessment of ion transport in soils during unsaturated flow. The parameter optimisation program PEST in combination with PHREEQC2 best simulated cation and anion transport in sandy soils under unsaturated conditions.