Through a combination of small scale laboratory experiments and access to measurements in a large stormwater tunnel, we have developed an understanding of processes that control the vertical rise of water in vertical ventilation shafts when large entrapped air pockets are released. Early laboratory investigations on the rapid filling of nearly horizontal pipelines indicated that strong upward surges of water could occur both as the leading edge of the air pocket arrived at the vertical shaft and as the trailing edge is expelled through the shaft and that the nature of the air/water interactions is fundamentally different in these two situations. It is also clear that small scale laboratory experiments are subject to significant scale effects and do not reproduce observations in actual large scale stormwater tunnels. This study presents modeling results that can reproduce the essential features of the air/water interactions in vertical shafts in laboratory studies that have been designed to separate the response to the leading and trailing edges of the air pocket. With modeling results, we can deduce which of the two phenomena are most relevant to the design of actual systems and what design variables are most critical to the formation of observed geysers. It is concluded that the entry of a large air pocket into a partially surcharged shaft is the critical process in the formation of observed geysers in actual systems and that small diameter shafts are more susceptible to geyser formation. Also, the water rise following the expulsion of the air pocket is basically an inertial surge process, only weakly dependent on the shaft diameter and strongly coupled to the local pressures that exist within the tunnel. These findings can be used to interpret the observed effect of a proposed geyser mitigation scheme that involves an increase in the shaft diameter above the tunnel crown.