Position C.2.b of Regulatory Guide 1.111 describes an approach to modeling the diffusion of effluents from roof top vents and short stacks using an elevated plume model under some conditions and using a ground-level source building wake model under other conditions. The approach is sometimes called a Split-H model. This report: presents the results of an evaluation of the technical basis for and utility of the concept behind the Split-H model, outlines the development of an upgraded Split-H model using existing plume rise and building wake models, compares ground-level, annual average concentrations estimated using the upgraded model with those estimated using the Regulatory Guide Split-H model and a ground-level building wake model, and discusses alternatives to the Regulatory Guide position that the NRC may wish to consider. Concentration comparisons are made using model results for meteorological data from 18 nuclear power plant sites.Differences between annual average, ground-level concentration estimates made using the upgraded Split-H model and those made with the other two models for the same meteorological conditions are a function of the distance from the source and stack parameters including exit velocity and diameter. They are greatest near the source. For example, at a distance of 400 m and a stack flow rate of 31.4 m3/s, the average ratio between the upgraded model and the corresponding ground-level release model concentrations ranges from about 0.6 for a stack exit vertical velocity of 5 m/s to about 0.2 for an exit velocity of 20 m/s. When compared with concentrations estimated using the Regulatory Guide Split-H model, upgraded Split-H model concentration estimates at 400 m range from about a factor of 5 larger for low vertical velocity releases to about a factor of 3 larger for high vertical velocity releases. Individual differences of an order of magnitude or more were found in both comparisons.On a climatological basis, the reduction of ground-level concentrations that results from the use of the upgraded Split-H model in place of the groundlevel release model is described by an exponential function of the ratio between the average wind speed in a sector and the vertical velocity of the release. The decay parameter in the relationship is a function of the distance from the source and the flow rate in the stack. It ranges from about -0.1 for long ranges (6400 m) and low flow rates (15.7 m3/s) to about -0.8 for short ranges (400 m) and high flow rates (125.7 m3/s). iii • ACKNOWLEDGMENTS Many people have contributed to the preparation of this report. The efforts of Clifford Glantz and the NRC Technical Monitor, James Fairobent are particularly noteworthy. Cliff assisted in the preparation of the data set used in the model comparisons, made the computer runs, and assisted in the analysis of the model output. Jim contributed to the work by providing encouragement and by highlighting the aspects of the work that are of particular interest to the NRC. Others, including