2. QG Model simulations and diagnoses
A series of experiments with the QG Model (Ramamurthy with modifications by Bramer/Mlodzik/Rose) were run with a baseline static stability and with a static stability estimated from the profiles of potential temperature during this period. The model is run with height data from the NCEP Eta model initializations, .5 degree horizontal and vertical resolution is centered at 42N, 83W, and has a domain size of 20 degs latitude by 30 degrees longitude. For the second set of experiments the static stability for the first 6 levels was decreased by 1/10th of the baseline values (average or climatology).
The model was run in diagnostic or "research" mode; attempts at forecasting mode were unsuccessful due to lack of convergence in the relaxation routines. The QG model does reproduce the gross features for the two evenings of the snow event. The closed 500 Mb low in the Great Lakes shifts slightly eastward during the two days and heights fall slightly from 527 DM on 11/10/96 at 00Z to 523 DM on 11/11/96. The strong speed shear near the base of the trough is also represented as are the low values of u and v winds near the broad center of the vortex.
Using the 3-dimensional arrays of omega and divergence from the QG model runs for 11/10/96 00z and 11/11/96 00z a series of cross-sections were produced. It was determined that a curved trajectory cross-section would be derived such that the cross-section included airstreams that moved over southern Lake Michigan and western Lake Erie. From upper level charts and radar it was clear that a lake-lake interaction was at play in the lake effect part of the storm beginning on 11/10/96; this trajectory did move over southern L. Michigan and the length of L. Erie. Because there was so little directional shear throughout the atmosphere during this period, the trajectory remained mostly unchanged with respect to height. The purpose of these experiments were to see if the lakes played any role in the large scale profiles of the QG model diagnoses. The a priori hypothesis is that if there is a response from the warm lake forcing on large scale omega and other dynamic fields, that response will be weak or undetectable. To produce a derived trajectory a simple asymptotic function was devised and parameterized to produce a curved path intersecting the two lakes:
where aspect and offset are tunable parameters, "lat" and "long" are base latitude and longitude. Both latitude and longitude are transformed to grid coordinates and the cross-section is produced from the latitude calculations.
Cross-section of omega along derived trajectory path determined to intersect Lake Superior, Southern L. Michigan and western L. Erie.
A cross-section of omega and divergence from the QG model is shown below. The strong short-wave which set off the heavy snowfall on the evening of 11/10/96 is represented by the weaker maxima extension at about 700 Mb near the grid point 25 in the y direction. This zone of ascent corresponds nicely with the eastern end of L. Erie the and northern quadrant of the short-wave, and matches with the area of best Q-vector convergence at 500 Mb.
Cross-section of Divergence (s-1) along derived trajectory path identical to previous image.
Note, that rising motions are contoured only while sinking motions are masked using the "fiddle" option of the software presentation tool. The strong maxima at each end of the domain correspond to jet stream dynamical motion in the strong shear zone of the polar vortex. Each strong maxima actually contains a sharp couplet of rising and sinking air that are reversed in the two cases. These likely indicate opposing quadrants of different jet streaks racing through the fast flow. The divergence field does not show much "knowledge" of the warm lakes.