Reverse osmosis is efficiently used for producing drinking water from groundwater sources containing dissolved impurities, including fluoride, ammonia, lithium, strontium, boron, arsenic, etc. The principal problems of utilizing reverse osmosis include scaling on membrane surfaces, concentrate discharges, and low permeate TDS that often require conditioning. The main goal of this work was to demonstrate the viability of a newly developed methodology that relies on low-rejection nanofiltration membranes to improve product water quality by increasing its TDS and calcium content, and its economic efficiency compared to conventionally used reverse osmosis. Disadvantages of employing reverse osmosis for the production of drinking water are attributed to the fact that several pollutants (including lithium, ammonia, and boron) are monovalent ions and, as such, are poorly rejected by membranes as compared to calcium, sodium, sulfate, and chloride ions. Thus, in cases in which lithium or ammonia are present in high concentrations, high rejection membranes are usually used that result in low TDS of the product water. This article presents the results of research aimed at developing a new approach to changing the ratio of monovalent and divalent ions in product water. The new method described in this paper relies on low rejection membranes in a two-stage application that enables us to reduce monovalent impurities and increase the concentration of calcium and TDS values in product water while leaving lithium concentration unchanged. This is achieved by applying a two-stage scheme with low-rejection membranes instead of the reverse osmosis stage. The two-stage treatment using nanofiltration membranes results in the same rejection of lithium and product water quality as reverse osmosis. However, the ratio value of calcium and lithium concentrations in the concentrate of nanofiltration membranes appears to be significantly higher compared with the ratio measured in the feed groundwater. This can be attributed to different rejections of these ions by membranes. Therefore, concentration (reduction of volume) of the feed water with nanofiltration membranes and further dilution of the concentrate with deionized water produce the same concentration of lithium and are associated with an increase of 2–4 times the concentration of calcium. Treatment of this water in the second nanofiltration membrane stage produces drinking-quality water with the required lithium content and increased calcium concentration. We focus on the real-world example of groundwater treatment in Yakutia, Russia, an area where lithium concentration exceeds drinking standards by 24 times. The paper presents a technique of ion separation and demonstrates experimental results that provide lithium removal while increasing the calcium concentration and TDS value. The resulting concentrations are 2–5 times lower than those obtained via conventional use of reverse osmosis membranes. A series of experiments were conducted to remove lithium from groundwater and demonstrate the efficiency of the newly developed method of ion separation. Experimental results of the concentration of obtained values of lithium, calcium, and TDS in permeate and concentrate flows at each membrane stage demonstrate that they provide separation of monovalent and divalent ions and increase product water TDS without increasing lithium. This experimental approach increases calcium and TDS values in product water by 2–4 times compared with the use of reverse osmosis membranes. Calculations of operational costs for different options (the use of reverse osmosis, two-stage nanofiltration, and ion separation in a two-stage approach) are presented. These results confirm the economic advantage of nanofiltration membrane applications to remove lithium as compared to the use of high-rejection reverse osmosis membranes. The increase in product water TDS facilitates the further reduction of concentrate flow rate and operational costs. The economic comparison involved the calculation of the required membrane area and number of membrane elements at each stage, calcium carbonate scaling rates, reagent consumption to prevent scaling, and amounts of concentrate discharged into the sewer. Experimentally obtained results confirmed the feasibility of increasing the calcium concentration and TDS values in product water by 2–5 times while leaving the lithium concentration at the same level. Design characteristics to calculate operational costs for conventional and new options are calculated and demonstrate a sufficient (30–40%) reduction of operational costs compared to conventional use of reverse osmosis. The reduction in reagent consumption is attributed to the utilization of low-rejection nanofiltration membranes that have lower scaling propensities compared with reverse osmosis membranes and a smaller payment for concentrate discharge. The developed approach to using two-stage nanofiltration instead of single-stage reverse osmosis provides multiple advantages that include improved product water quality, lower concentrate consumption, and lower reagent consumption that are attributable to the use of low-rejection membranes. Different case studies are planned to demonstrate the efficiency of the proposed techniques to reduce ammonia, fluoride, and boron in drinking water. [ABSTRACT FROM AUTHOR]