[en] The Rocky Flats Environmental Technology Site (RFETS) is owned by the U.S. Department of Energy and began operation in 1952 as a nuclear weapons production, research, and development facility. The plant closed operations in 1989 and is currently being decommissioned and cleaned up. Significant portions of the land will ultimately be made available for public use. The plant is located on 2650 ha of property about 26 km northwest of downtown Denver, Colorado. The primary residual radionuclide of concern in soil is plutonium 239/240. Most of the material is located near the soil surface, and was deposited during the late 1960s as a result of the resuspension of contaminated soil from natural wind erosion and man-made resuspension during remediation activities. This paper reviews the results of a project to develop soil action levels in activity per gram of soil that are based on prescribed dose levels for the public. The soil action levels will be used for decision-making during cleanup and remediation of the site and would prevent people from receiving doses greater than an established limit. The project has been conducted with oversight from a panel made up of members of the local community and government officials. The panel provided significant input regarding the selection of parameters and characterization of people who might inhabit the land in the future. Following review of a number of available computer codes, the RESRAD code was selected as the basis for the calculation. Some features of RESRAD were not adaptable to site-specific data and modifications were needed. Prior to making calculations, a sensitivity analysis was conducted to determine the most important parameters. Inhalation of resuspended soil was a key pathway influencing the soil action levels and therefore site specific data were incorporated to the greatest extent possible in characterizing this pathway. Resuspension factors for the site were based on historical measurements of plutonium in soil and air. The code was also modified to incorporate a feature that takes into account the possibility of a prairie grass fire that could significantly enhance resuspension. A series of scenarios were developed typical of people who now live in the area and who would be most likely to receive the highest dose based on their occupation, diet, and lifestyle. Distributions of soil action levels were developed of activity in soil as a function of probability of exceeding the established dose limits for the different scenarios. Soil action levels were proposed to the panel based on these distributions. The project represents a major step forward in the development of clean up standards for contaminated soil not only because of the use of available site-specific data and estimation of uncertainties, but also because of the open, public process that was used. (author)