334 February 38.7 21.3 17.4 348480 4.03 March 42.1 23.2 18.9 362880 4.20 April 15.2 8.4 6.8 259200 3.00 May 8.9 4.9 4 172800 2.00 June 5.1 2.8 2.3 86400 1.00 Reactive transport was performed with flux-flux boundary conditions, 50 cells, one shift per month, and conservative mixing/dispersion parameters (cell length 1 m, low longitudinal dispersivity, and explicit molecular diffusion) to avoid excessive numerical dispersion and favor the geochemical signal of hydrological forcing. The outputs were extracted with SELECTED_OUTPUT (pH, pe, alkalinity, equilibrium phases, and kinetic reactants) and USER_PUNCH, which converts total concentrations to mg/L by multiplying by molecular weights, maintaining PHREEQC's internal mass control. To simulate the monthly evolution of reactive mineralogy over multidecadal time horizons, the sequential month-by-month run was automated using Python + phreeqpy/IPhreeqc. In each iteration, the script executed the file for the month, read the SELECTED_OUTPUT, identified the remaining moles of each kinetic reagent (columns with the prefix k_ in the PHREEQC code output), updated the next file by replacing the -m values, and adjusted the -time_step according to an external table (Excel) of monthly contact times. This cycle was repeated for each scenario, allowing the progressive depletion of reactants and monthly climate forcing to control the system response in a coupled manner without coupling external flow models. Each monthly infiltration pulse mixes a fraction of meteoric water with the resident water in the package using a sequential mixing scheme, followed by oxygencontrolled reactions and subsequent drainage of the system. This sequence was modeled as a one-dimensional column, where changes in water quality and mineral transformation are evaluated month by month. Figure 1 schematically shows this conceptual dynamic applied to the simulated system.
RkJQdWJsaXNoZXIy MTM0Mzk2