AN EXAMINATION OF BREAKWATER
PERFORMANCE AT BURNS HARBOR, INDIANA
James P. McKinney and Margaret A. Sabol
U.S. Army Corps of Engineers Engineering Research & Development Center
Vicksburg, MS, 39180
July 15, 2003
This discussion compares incident and transferred spectral results for gages located
at Burns Harbor, Indiana. Comparisons are made by examining the differences in the
incident and transferred harbor energy spectrums. Spectral analysis allows the energy
of the total wave record to be broken down into discrete frequency bands. Energy inside
and outside the harbors may then be compared and a transfer factor for each discrete
frequency can be determined.
Incident and harbor wave data were collected within and offshore of the breakwater
at Burns Harbor, Indiana. The purpose of this data collection effort was to determine
characteristics of the rubble mound breakwater located there. Wave records were col-
lected hourly using subsurface pressure sensors. The sample rate for these sensors was
1 Hz and the burst length was 2048 seconds.
The analysis utilized the Welch, [1], spectral analysis method with 50% overlapping
segments. Since the raw time series were obtained using sub-surface systems, a depth
determined high frequency cutoff was applied. The averaged co-and quad-spectra from
each analyzed record were used to calculate significant wave height (Hm0), peak period
(Tp), and mean wave direction at Tp (Dp) and energy spectrums.
The time period 6 January 2003 0700 thru 7 January 2003 0 was selected for de-
tailed analyses. Figure 1, top plot, shows six successive hourly energy spectrums for
IN001, located in the open lake. The lake is relatively flat at 0700 with Hm0 0.53 m
and Tp about 4.3 seconds. Over the next 5 hours the Hm0 builds to 1.86 m and the
Tp moves to 6.6 seconds. The energy at the long period end of the spectrum is very
small. Examining simultaneous analysis results from within the harbor provides a de-
scription of the breakwater's performance. Figure 1, bottom plot, shows spectrums for
the same six hours at the inside gage, IN002. There is almost no energy at 0700 GMT
1/6/03, with Hm0 0.05m and Tp 4.5 sec. In subsequent hours, bimodal spectrums de-
velop as more and more energy is transferred through the breakwater, with peaks >20
seconds and between 6 and 10 seconds. The magnitudes of these peaks are small when
compared to energy for those frequencies outside the harbor.
To provide a more direct comparison of incident and transferred energy, a transfer
coefficient (xfer) can be calculated by dividing the transferred energy at each frequency
by the corresponding outside energy, eqn.1.
1