Soil hydraulic and solute transport parameters are crucial for characterizing subsurface hydrological processes, yet their accurate identification in deep soils (> 5 m) remains challenging in regional-scale studies. The inversion estimation of these parameters with available soil indicators is a good alternative method to aid potential groundwater recharge (i.e., downward water flux from the bottom of root zone toward the aquifers) estimation. Here, we collected over 3000 soil samples from 39 profiles > 10 m deep, with varying climate, vegetation and soil conditions, to determine the texture, organic carbon and solute contents (chloride, nitrate, and tritium) in the Loess Plateau of China. The soil hydraulic (soil water retention curve and saturated hydraulic conductivity) and solute transport parameters (dispersion coefficient and migration velocity) were then calculated using multiple Pedotransfer functions and modified breakthrough curve method. These parameters were used to estimate potential groundwater recharge with the Darcy's law and tritium/chloride/nitrate peak methods, thereby discussing their uncertainties in soil parametrization. Results showed that the soil properties and hydraulic parameters were more variable with locations rather than land use types. The coarser soils (sand of 42.0 +/- 8.3 %) exhibited higher saturated hydraulic conductivity (38.2 +/- 27.4 cm d(-1)) in the northern regions. The parabolic chloride and nitrate profiles yielded dispersion coefficients of 0.07-0.33 m(2) yr(-1) and migration velocities of 0.07-0.20 m yr(-1), with higher values in the northern regions - suggesting soil texture as the key driver of solute transport. The estimated potential recharge rates ranged 0-69.9 mm yr(-1), with higher values observed in the northern regions (17.8 +/- 16.3 mm yr(-1)) and under shallow-rooted plants (22.4 +/- 14.4 mm yr(-1)), attributed to the coarser soil texture and limited plant transpiration effects, respectively. The uncertainties in the hydraulic parameter-based methods stemmed from simplified assumptions and parameter estimation, while the solute-based methods were affected by timescales, plant absorption and biogeochemical cycles. This study provides technical support and essential datasets for model parameterization within deep vadose zones.