The Picnic Rocks Channel Current
Figure 1: Rocks responsible for increasing channel current speeds at Picnic Rocks. Image taken from Googlemaps.
What Makes This Channel Current so Dangerous?
The channel current at Picnic Rocks is especially dangerous. The location is just offshore from Marquette, MI, which is the largest city in the Upper Peninsula. This means that a greater number of people have fairly easy access to the site, increasing the likelihood of an individual encountering the potentially life threatening channel current. Another reason that the Picnic Rocks channel current is especially dangerous (as compared to other areas) is not only due to the fact that the current speeds are increased between the shore and the off shore obstacle(s), shown in Figure 1, but also due to the fact that water moving into this channel is also forced up and over a sandbar about 4 to 5 feet below the water surface. The combination of these two forcing mechanisms can cause the channel current to increase to speeds at or above 2 mph very rapidly. Current speeds do not necessarily need to be 2 mph to be dangerous as slower currents, around 0.5 mph, can cause a novice swimmer to struggle significantly. It is very difficult at this point to determine what is considered dangerous as these values vary from person to person, depending upon swimming experience. After discussing the issue with several experts around the Great Lakes, it was determined that 0.5 mph may be the best value to use for a moderate risk to swimmers.
Picnic Rocks Channel Current Data Review (2011-2012)
|Figure 2: Acoustical Doppler Current meter similar to the one used at the Picnic Rocks location.||Figure 3: Shows the 5 layers labeled Cell 1 through Cell 5 and the directions that the current meter samples.|
How Did We Collect All of This Data?
Data acquisition at Picnic Rocks occurred between June 14, 2011 and October 24, 2011 and between June 1, 2012 and October 3, 2012 by using an Acoustical Doppler Current Meter like the one shown in Figure 2. The meter was positioned between the rocks off the shore of Marquette and the shoreline, but it was not placed directly above the sandbar, which could cause values to be decreased to some extent. The current meter measures data at 5 different layers as shown in Figure 3, with each layer being approximately 3 feet deep, except for the top layer which varied greatly depending on the lake levels. To keep things simple, the current meter works by detecting the velocity of particles flowing past the meter in each layer. Data was not only collected from the current meter, but it was also collected from nearby surface observation systems that collect wind data. Both datasets were extremely important in the completion of this analysis and will be expanded upon in the next section.
What Data Analysis Methods Were Used?
Initially, there were two main areas of focus in this study. The first area of focus was to determine if there was a correlation of a specific wind direction with increased channel current speeds using data from the Marquette Coast Guard station, the NMU observation site, as well as from the current meter. Ideally, a correlation between the wind direction and increased channel current speeds would allow forecasters to give advanced notice that channel current development is possible when winds are expected to be from specific directions.
The second area of focus was to determine if a correlation exists between increased channel current speeds and increased wind speeds. It was expected that higher wind speeds would help to push the water through the channel at a much greater speed causing the channel current to develop and strengthen. These data were compared for each of the previously mentioned observation platforms and a comparison was also completed between the 2011 and 2012 data. After the first year of study, it was decided that analyzing the wave height, direction and period may yield a better correlation with increased channel current speeds than analyzing the winds. There is very little observed wave data near the Picnic Rocks location, so it was decided to use modeled wave data provided by the Great Lakes Environmental Research Laboratory (GLERL). The wave data was only archived for the 2012 season at this point and will be archived again in 2013 for comparison.
What Did We Find?
To begin with, the entire dataset was analyzed to see how the channel current speeds were distributed. To do this it was necessary to classify each channel current speed into bins of 0.25 mph increments and then total the number of events that occurred in each bin as shown in the graph below:
It can be seen that most of the events that occurred were 1 mph or lower, while around 40 events or less occurred with channel currents speeds greater than 0.75 mph. It is worth noting, for completeness, that a few events in 2011 were greater than 1.75 mph and were likely due to the calibration process immediately after the Acoustical Doppler Current Meter was installed. The comparison between the 2011 data (Figure 4) and 2012 data (Figure 5) show very similar results in most of the classifications. The only major difference is in the 0.25 mph classification, where events in 2012 were nearly twice the 2011 data. As mentioned in previous sections, the channel current speeds did not exceed what is expected to be dangerous for an experienced swimmer (2 mph), but the currents were high enough to potentially cause problems for a novice swimmer (>0.3 mph). Channel current speeds were expected to be greater than what is shown in the chart and it is possible that the speeds are slightly reduced due to the equipment being placed away from the sand bar. After exploring the possibility of adjusting the siting position of the current meter it was determined that siting was limited due to communication issues as well as water depth issues above the sand bar.
Did We Find a Correlation Between Wind Direction and Increased Channel Current Speeds?
After analyzing the data distribution it was then necessary to take a closer look to see if wind direction played a role in the development of channel currents greater than 0.3 mph. In order to ensure that a predominant wind direction did not erroneously lead to a higher correlation, it was necessary to normalize the wind events associated with greater than 0.3 mph channel current speeds to the total number of wind events from each direction. Once this was complete, it was then much easier to see that north and south events maintained a greater correlation than most of the other directions; however, both site also show an increased correlation from the northwest and a slight increase from the southeast as shown in Figure 6. This is reasonable as the obstacle(s) and channel have a slight northwest orientation as well. Layer 3, which is the next three feet below Layer 4, showed similar results, but fewer overall channel current speeds exceeded 0.3 mph in that layer, especially in the 2011 data. All the layers below Layer 3 had one percent or less channel current speeds greater than 0.3 mph.
Did the Wind Direction Correlate to Increased Currents in 2011 and 2012 for the Coast Guard Site?
|Figure 7: Wind direction associated with increased channel current speeds normalized to percent of total winds from each direction for each site using Layer 4 channel current values for 2011 and 2012 data collection.|
Did the Wind Speeds Correlate to Increased Currents?
|Figure 8: Percentage of increased channel current speeds, in Layer 4, from each wind direction that were associated with wind speeds less than 10 knots (11.5mph) from the MCGM4 and KNMU sites. These data were collect during the 2011 season.|
Once the wind directions were analyzed it was hypothesized that greater wind speeds may be associated with increased channel current speeds. The graphs in Figure 8 show the distribution of winds speeds less than 10 knots (11.5 mph), which were normalized to percent of total winds from each specific direction. It can be seen that a majority of the increased channel current speeds are associated with wind speeds less than 10 knots. These results were seen in Layer 3 as well. One theory is that, as the winds increase to 15 knots (17 mph) or higher, a greater deal of the energy is applied to wave generation as opposed to current generation. At this point it is still very difficult to forecast channel current development based on the data shown; however, studies are currently underway that will potentially give us a better understanding about channel current development.
|Figure 9: Comparison between the 2011 and 2012 data from the MCGM4 site. The graphs show percentages of increased channel current speeds, in Layer 4, from each wind direction that were associated with wind speeds less than 10 knots (11.5mph).|
When comparing the 2011 and 2012 wind speed data, as shown in Figure 9, similar results were found. The 2011 data shows 90 to 100 percent of the events from each direction were associated with wind speeds less than 10 knots, while the 2012 data shows slightly less of a correlation with only 70 to 80 percent of the increased channel current events associated with wind speeds less than 10 knots. This makes it very difficult to use these data for forecasting channel current development as winds through the swimming season are generally less than 10 knots. Using the wind data alone would create conditions favorable for over-forecasting channel current development.
What did the Wave Data Show?
| Figure 10: Total channel current observations from each direction and for each wave height from the 2012 dataset. The red selections indicate areas of significantly increased channel current observations.
| Figure 11: Percentage of the total observations from each direction and each wave height that were associated with channel current speeds greater than or equal to 0.3 mph. The red selections indicate percentages greater than 60 as well as the wave range that is better correlated with increased channel current speeds.
The archived wave data associated with each wave direction can be seen in Figure 10. The interesting points, highlighted in red, indicate that the increased number of channel current observations occurred from the south and the northwest. This seems logical due to the northwest to southeast orientation of the shoreline and rock structures in the Picnic Rocks area. The next step was to analyze the percentage of the total observations from each direction and each wave height that were associated with channel current speeds greater than or equal to 0.3 mph as shown in Figure 11. The wave direction and wave period that were 60 to 100 percent correlated with increased channel current speeds were highlighted in red. The reason these events are of greater interest is that the false alarm rates would be significantly lower for the channel current forecast. In the end it appears as if wave heights of 2 to 5 feet from the north, northwest or with a southerly component show a better correlation with increased channel current speeds. The data analysis for the 2011 data will be available soon.