A waterspout funnel and spray ring were observed under a cumulus line over the Great Salt Lake for about 5 min shortly after sunrise on 26 June 1985. Videotaped features strongly suggested that the funnel rotation was anticyclonic, These observations have been used as the basis for a study of the initiation and evolution of waterspouts through a series of numerical experiments at two scales, that of a cloud and a waterspout.
The cloud scale has been simulated using an improved Goddard-Schlesinger model with nearby Salt Lake City soundings. The main model improvements have been 1) a parameterized, three-class ice phase and 2) a line initialization in addition to the more common axisymmetric buoyant bubble. Cloud-scale vortex pairs developed for each mode of initiation, but a much stronger, more upright, low-level anticyclonic vortex grew from the line initiation than from the bubble. However, cumulus-scale vortices are common while waterspouts are rare, and the real test of a model is whether a waterspout can develop in the limited cumulus lifetime.
The 600-m horizontal grid of the cloud model cannot resolve waterspouts, and a modified Monash high-resolution axisymmetric vortex model with vertical domain and small section has been “embedded” at selected positions and initiated at selected times in the computed flow field of the cloud. Many experiments have been carried out with the vortex model. In the most important series, the boundary conditions were changed with the fields of the model cumulus as it evolved, and the time at which the vortex was started was varied through the lifetime of the parent cloud. Results showed that for each mode of cloud initiation, the vortex that started at the anticyclonic center grew faster than those started at other centers. This result fits with the observed anticyclonic rotation of the waterspout, strongly suggesting that the cloud vorticity was important in its initiation. The greatest azimuthal speed for the bubble-initiated cloud was 11 ms−1 when the vortex model was started at 28 min cloud time with time-varying boundary conditions, whereas it was 21 m s−1 when started at 12 min in the line-initiated cloud. Speeds were comparable when the inner domain moved with the anticyclonic cloud center. These speeds are close to the spray-ring threshold azimuthal velocity of roughly 22 m s−1 estimated by Golden from photographs.
Together, these model results support the hypothesis that, at least in some circumstances, cloud processes alone can produce waterspouts in the absence of external vorticity sources such as surface convergence lines or other shear features.
† Present affiliation: WSFO/National Weather Service, NOAA, Topeka, Kansas.
* Present affiliation: Department of Meteorology, The Pennsylvania State University, University Park, PA.