The chloride-mass-balance (CMB) method has been used extensively to estimate recharge in arid and semi-arid environments. Required data include estimates of annual precipitation, total chloride input (from dry fallout and precipitation), and pore-water chloride concentrations. Typically, CMB has been used to estimate ancient recharge but recharge from recent land-use change has also been documented. Recharge rates below a few mm/yr are reliably detected with CMB; however, estimates above a few mm/yr appear to be less reliable. We tested the CMB method against 26 years of drainage from a 7.6-m-deep lysimeter at a simulated waste-burial ground, located on the Department of Energy s Hanford Site in southeastern Washington State, USA where land-use change has increased recharge rates. Measured drainage from the lysimeter for the past 26 years averaged 62 mm/yr. Precipitation averaged 190 mm/yr with an estimated chloride input of 0.225 mg/L. Initial pore-water chloride concentration was 88 mg/L and decreased to about 6 mg/L after 26 years, while the drainage water decreased to less than 1 mg/L. A recharge estimate made using chloride concentrations in drain water was within 20 percent of the measured drainage rate. In contrast, recharge estimates using 1:1 (water: soil) extracts were lower than actual by factors ranging from 2 to 8 or more. The results suggest that when recharge is above a few mm/yr, soil water extracts can lead to unreliable estimates of recharge. For conditions of elevated recharge, direct sampling of pore water is the preferred method, because chloride concentrations are often 20 to 50 times higher in directly-sampled pore water than in pore-water extracts.