Implications of Small-Scale Flow Features to Modeling Dispersion over Complex Terrain

Abstract

Small-scale, topographically forced wind systems often have a strong influence on flow over complex terrain. A problem is that these systems are very difficult to measure, because of their limited spatial and temporal extent. They can be important, however, in the atmospheric transport of hazardous materials. For example, a nocturnal exit jet-a narrow stream of cold air-which flowed from Eldorado Canyon at the interface between the Rocky Mountains and the Colorado plains near the Rocky Flats Plant (RFP), swept over RFP for about 3 h in the middle of the night of 4?5 February 1991. It extended in depth from a few tens of meters to approximately 800 m above the ground. Because the jet was so narrow (2 km wide), it was poorly sampled by the meteorological surface mesonet, but it did prove to have an effect on the dispersion of tracer material released from RFP, producing a secondary peak in measured concentration to the southeast of RFP. The existence and behavior of the jet was documented by Environment Technology Laboratoy's Doppler lidar system, a scanning, active remote-sensing system that provides fine-resolution wind measurements. The lidar was deployed as a part of a wintertime study of flow and dispersion in the RFP vicinity during February 1993. The MATHEW-ADPIC atmospheric dispersion model was run using the case study data from this night. It consists of three major modules: an interpolation scheme; MATHEW, a diagnostic wind-flow algorithm that calculates a mass-consistent interpolated flow; and ADPIC, a diffusion algorithm. The model did an adequate job of representing the main lobe of the tracer transport, but the secondary lobe resulting from the Eldorado Canyon exit jet was absent from the model result. Because the jet was not adequately represented in the input data, it did not appear in the modeled wind field. Thus, the effects of the jet on the transport of tracer material were not properly simulated by the diagnostic model.