Abstract
Land use change from natural ecosystems to cropland influences groundwater recharge, including water quantity and quality. Soil core samples (0–11-m depth) from six boreholes beneath irrigated cropland and two boreholes beneath natural ecosystems, in Vertisols, were analyzed for particle size distribution, water content, and water-extractable Cl−. Chloride mass balance and numerical, one-dimensional unsaturated flow and transport modeling were used to assess average and transient recharge fluxes and to test matrix vs. preferential flow hypotheses. Water contents under irrigated cropland were significantly higher than under natural land with similar particle size distributions. Pore-water Cl− concentrations in deep vadose zones (>3 m) under irrigated cropland (900–2000 mg L−1) were similar to recent local groundwater Cl− and significantly lower than pore-water Cl− in deep vadose zones under natural land (3000–6000 mg L−1). Calibrated models' recharge rates through the soil matrix were much higher under irrigated cropland (90–230 mm yr−1) than natural ecosystems (1–3 mm yr−1) and were consistent with groundwater-balance estimates of average recharge (110–160 mm yr−1). In contrast, matrix recharge rates under natural ecosystems were considerably lower than those based groundwater balance (50–80 mm yr−1). While matrix unsaturated flow under irrigated cropland explains vadose zone and groundwater observations, under natural ecosystems preferential flow paths are required to support observations. Plowing and irrigation prevent development of crack networks and promote matrix percolation through the clay, which flushes salts from previously immobile vadose zone pore water. These phenomena may be applicable to similar land use changes in Vertisols globally.