(h) Test Procedures . A model Stillwater level should be 

 selected that closely approximates the higher water stages that normally 

 prevail during severe storms in the prototype. This requires the study 

 of water level records in the prototype locality, with proper attention 

 given to the higher levels experienced in the area in the past. Water 

 levels in Vermilion Harbor are determined by the water surface elevation 

 of Lake Erie, which can vary monthly and also yearly. These seasonal and 

 annual variations in the water level are related to volumetric changes in 

 the lake, which are principally caused by precipitation, evaporation, and 

 runoff. The usual pattern of seasonal variation of water levels in the 

 Great Lakes consists of highs in summer and lows in late winter. The 

 highest monthly average level is usually recorded in June and the lowest 

 in February, though seasonal fluctuations have occasionally caused no- 

 ticeable departures from this pattern (U.S. Army Engineer District, Lake 

 Survey, 1952). IVind setup and seiches, which can cause the water level 

 at a particular locality to vary daily and also hourly, are relatively 

 short -period fluctuations superimposed on the longer variations in lake 

 levels. These short-period oscillations are due to a tilting of the lake 

 surface generally caused by wind and possibly by differentials in baro- 

 metric pressure. Large short-period rises in local water level are asso- 

 ciated with the most severe storms which generally occur in the winter 

 months when the lake level is usually low; thus, the probability of a 

 high lake level and a large wind setup or seiche occurring simultaneously 

 is relatively small. A Stillwater level of +3.0 low water datum (LWD) was 

 selected for use in the model study. This value was determined by combin- 

 ing the average water level of Lake Erie during the ice-free period (+2.1 

 feet LWD) with an assumed 0.9-foot short-period rise in local water level 

 due to wind setup. The entrance to Vermilion Harbor is exposed in vary- 

 ing degrees to storm waves from directions ranging clockwise from west to 

 northeast. The most severe wave action at the harbor entrance is caused 

 by storms from the north and northeast. Measured wave data on which to 

 base a comprehensive statistical analysis of wave conditions were not 

 available for the Vermilion area; however, records were available from 

 which statistical wave hindcast data could be compiled. These records 

 were U.S. Coast Guard wind data for a 6-year period (1946-1951) at Lorain 

 Harbor, Ohio (11 miles east of Vermilion Harbor), and it was assumed that 

 winds with similar characteristics could be expected to occur at Vermilion. 

 The frequency of occurrence of winds of varying velocities from the dif- 

 ferent directions of storm approach has been compiled for the normal navi- 

 gation season on Lake Erie (April to November). The characteristics of 

 waves that could be expected to approach Vermilion Harbor were determined 

 by applying these wind data and the fetch lengths corresponding to each 

 direction to the theory of wave hindcasting. The deepwater wave charac- 

 teristics were converted to shallow-water values, at the position of the 

 model wave generator, by wave refraction techniques. Wave height, period, 

 and direction characteristics used in the model tests (selected from the 

 wave data obtained as previously discussed) are presented in Table 4-7. 

 The data obtained during the testing program included (a) wave height 

 measurements at several selected locations throughout the harbor area 

 and the river channel, (b) photos of wave patterns, and (c) visual obser- 

 vations of wave action in the model. Wave heights measured in the model 



256 



