FLUCTUATIONS OF WATER LEVEL IN A CANAL CAUSED BY THE FILLING OF LOCKS

BY

Mr. LOUIS C. SABIN
Assistant Engineer

The two locks now in operation at St. Marys Falls Canal, Mich., are located side by side, connecting with the same channel. The Weitzel Lock is 80 feet wide in the chamber, 60 feet at the gates, and the length between hollow quoins of the service gates is 515 feet. The Poe Lock is 100 feet wide throughout, and 800 feet in length between hollow quoins of service gates. The lift is, ordinarily, about 19 feet. The locks are filled by culverts under the floor, closed by butterfly valves under the upper miter wall; the culverts are 8 by 8 feet in size, and the valves 8 by 10 feet, giving a clear opening at the valve of about 8 by 81⁄2 feet. There are two culverts in the Weitzel Lock and six in the Poe Lock.

The upper reach of the canal is about a mile in length from its junction with the river to the head of the locks. It was excavated in earth and rock to a depth of 25 feet below low water, the rock sides being left rough and the canal face above the rock being formed of timber cribs. In the work of widening the canal, recently completed, the rock along the north side was channeled and concrete revetment walls built on top of the rock throughout the greater part of the length, only the upper end of the north wall being of timber crib construction.

When the valves are opened to fill one or both of the locks the water level in the canal above the gates is depressed, the extent of the lowering depending upon the number of valves opened. This depression of the water surface at the lower end sets up a current in the canal, which continues to flow downstream until the surface above the lock is about as far above the normal level as it had been depressed below the normal by the sudden opening of the valves. The current then starts back upstream, resulting in again lowering the water in the canal. If the water is not disturbed by again opening the valves, these fluctuations of the water surface

continue, with great regularity, for some time after the valves are closed.

Under conditions formerly existing, the changes in velocity and the reversal of the direction of flow in the canal made navigation difficult and unsafe, especially in view of the fact that the boats must pass in and out through one opening having a width of only about 108 feet. To remedy this condition, as well as to give additional room for navigation, the canal has recently been widened, and the effect of this increase in cross-section upon the currents and fluctuations in water levels has been determined.

A sketch plan of the canal is shown on Plate I. The full line shows the present outlines of the canal, while the dotted line shows the northern limit as it existed prior to the widening.

The fluctuations of water level in the canal caused by filling the locks are shown in Plates II and III. The increase in amplitude of the fluctuations, as the number of valves opened is increased, is brought out in Plate II, where the rate of filling the Poe Lock is also shown in some cases. It may be said, however, that the amplitude does not depend, alone, upon the number of valves opened, but also upon the phase of the wave, or the condition of the water level, whether rising or falling, at the time the valves are opened; in other words, to open four valves when the water level has begun to fall may cause a greater fluctuation than to open six valves when the water level is rising. If the valves were always opened at a time when the water level is constant, the effect of the number of valves on the amplitude of fluctuation would be more clearly brought out, but the locks are seldom out of use long enough to permit the water in the canal to come to rest. The observations on Plate III indicate the decrease in amplitude with the increase in distance from the locks, the fluctuations becoming practically nil at the upper end of the canal where it joins the river. It will be noticed that the rise or fall at the different points along the canal is almost simultaneous, and that the time length of the wave, or the interval between two successive high water stages, is practically constant at twelve to thirteen minutes, without regard to the number of valves opened. Part of the observations shown on Plate III were made prior, and part subsequent, to the widening of the canal, and the effect of the enlargement of the channel upon the character of the fluctuations may thus be seen. It appears that after the widening the time between two successive high water stages remains the same as before, but the amplitude is considerably diminished.

The most important effect of the widening is the reduction in the velocity of the current in the restricted section of the canal, as it is this, rather than the rise and fall of the surface, that interferes with navigation. By considering the amount of water drawn off into the lock, and the amount required to fill the canal from a low stage to a higher one, the velocity of the current through the restricted section may be obtained.

Comparing in this way the flow through the restricted section. at 700 E. (see Plate I) for five minutes (10.22-10.27 a. m., November 20, 1903) with that for five minutes (11.24-11.29 a. m., October 18, 1910), while the Poe Lock was being filled with six valves in both cases, the following results are obtained:

The following conclusions may be drawn from the observations: The fluctuations caused by the filling of the locks, before the widening of the canal had an amplitude of 2% to 5 feet.

The time interval between successive high water stages is twelve to thirteen minutes.

From one to two hours time is required for the water in the canal to come to rest after an initial disturbance of considerable magnitude.

The amplitude of the fluctuations depends upon the number of valves opened, the sign of the change in stage of water in the canal at the time of opening the valves-that is, whether rising or falling and on the cross-section of the canal.

The water level reaches the highest stage at nearly the same time throughout the length of the canal, but the amplitude decreases gradually from the locks to the entrance, being nearly zero at the latter point.

The enlargement of the canal reduced the amplitude of the fluctuations by about 50 per cent, and the velocity at the restricted section by about 60 per cent, but had no effect on the time interval between successive high stages.