Tuesday, December 17, 2013

Sediment Transport of Lester River during Duluth's June 2012 Flood

Sediment transport of Lester River during Duluth’s June 2012 Flood

Photo Credit: Bob King

Text Box: Photo taken by Bob King

Bob Kryzer
Jeffery Harrison
Shane Loeffler

The June 2012 flood was a disastrous weather event that occurred from Tuesday, June 19th, through Wednesday afternoon on the 20th. In the week prior to the storm there had been moderate rainfall events that had saturated the ground. Reports from Accuweather.com show that the Duluth Airport had recorded 1.76 inches of rainfall on June 14th and another .95 inches on June 17th. This saturation caused the heavy rainfall on the 19th to no longer soak into the ground, quickly raising the levels in drainage ditches, creeks, and rivers due to increased overland flow.
The weather event was set up by a large body of warm moist air being overridden by a cold front. The storm built over northeastern Minnesota with a direction of travel such that the large, slow moving thunderstorms tracked over the paths of each other. This generated a combined effect resulting in a huge quantity of precipitation. The post storm reports showed that rainfall ranging from 8-10 inches fell over a large area in 24 hours (Graning, 2012).
            Lester River at the eastern end of Duluth developed a particularly impressive mouth bar which showed up as an arc 55 meters from the normal terminus of the river. The remnants of the bar are still visible today.
            The Lester River and Amity Creek watershed drains a total area of 52.1 mi2.  This large area, combined with the enormity of this five-hundred year flood event ("Flooding Rains in Northeast Minnesota: June 19-20, 2012.") and the steep slope of this river led to a massive mobilization of sediment and subsequent deposition.  This deposition accumulated most dramatically at the mouth of the river as a large offshore mouth bar.  Having been reworked by a year and a half’s worth of wave action, the bar is now located near the shore and is accessible by foot from London Road.  Photographic and video documentation of the flood, as well as high stand debris markers, allow us to reconstruct flow conditions during this rare event. Using this documentation, field and remote measurements, and an Excel Software Plug-in we will present a reasonable measure of peak discharge needed to create the mouth bar seen today.  

            Cross sections of the river were taken using stadia rod in conjunction with a laser range finder and an eye level. Two locations were utilized, 1) on the upstream side parallel to London Road Bridge, and 2) about 150 meters up river from the bridge. Due to the cross section near the bridge being of low resolution due to our difficulties with the kayak data acquisition method it was not used in our final calculations.  
The gauging station at Amity Creek was destroyed during the flood so we used two approaches utilizing comparative analysis to find the discharge. A Post-it Note reference we gained from personal communication with Karen Gran lead us to a discharge rate of 2070 cubic feet per second (cfs) for Amity Creek. Through comparing this discharge value and Amity’s drainage area (16.5 mi2) to the watershed area of Lester River (35.6 mi2) we determined an estimated discharge rate for Lester River to be 4460 cfs, for an estimated combined flow of 6000 cfs. We also compared the total drainage area of both Lester and Amity (52.1 mi2) to the watershed size of the Knife River along with an estimated discharge as determined by the USGS for the Knife River (~25,000 cfs; Czuba et. al., 2012). With these approaches, an upper bound of 16,000 cfs was found along with a lower bound of 6000 cfs.
To measure the estimated volume of the deposited mouth bar, our study group conducted another survey. Using a stadia rod, laser range finder, and eye level, we measured a profile on the riverside adjacent to the mouth bar, a profile across the crest of the mouth bar, and a longitudinal profile over the mouth bar. It was easy to tell what clasts are associated with the reworked mouth bar and those of the lake shore due to their differing angular characteristics. Using the average river depth and average crest height we determined an estimated cross sectional area and multiplied this with the length of the mouth bar to determine an estimated volume of deposited material.
To fully determine the sediment transport rate of Lester River during the flood, we need more data. A grain size distribution was conducted using a Wolman pebble count with 116 clasts (Wolman, 1954). The surface slope of the reach from the delta to upstream survey site was determined using LiDAR and determined to be 0.007. Lastly, a Manning’s n roughness value was estimated (0.04) based on accepted values for a gravel bedded stream (Chow, 1959).
Data processing was done by BAGS (Bedload Assessment in Gravel-bedded streams; Pitlick et. al., 2009), a Microsoft Excel extension. BAGS accepts inputs of discharge rate, river profile, and grain size cumulative percentage, and outputs a bedload transport and transport capabilities. BAGS also give the height of flow needed to sustain such an inputted discharge rate for a given cross-sectional area. Additional inputs into BAGS were surface slope and an estimated Manning’s n.

            The river profile parallel to the London Road Bridge was not used in the investigation as we determined it to be of too low of a resolution. We used the collected channel geometry from 150 meters upstream of the bridge. This profile is illustrated in Figure 1 and values are shown in Table 1. Flood debris was observed in trees about 4.2 meters above baseflow.
Figure 1 - Estimated scatter plot of collected channel geometry. Blue line denotes the land surface, red line denotes baseflow conditions, and the green illustrates 2012 flood level as determined by flood debris observation. Data collected in field consists of distances 0-43.4m, the rest of the points are estimated to contain the 2012 flood level with values from Google Earth.
Table 1- Values collected during channel geometry survey. Elevations are corrected to water surface elevation at the time of survey. (*) Symbol denotes estimated values entered to contain the 2012 flood within the channel geometry.
Elevation (m)
Range (m)
Elevation (m)
Range (m)
Grain size analysis of 116 clasts from the upstream survey site revealed a D50 of 64mm, shown in Figure 2.
Figure 2- Cumulative grain size distribution of Lester River in vicinity of geometry survey using Wolman pebble count.
            Utilizing BAGS and the upper discharge bound of 16,000 cfs, a total of approximately 6,500 kg/min of sediment was found to be moving through the system.  The lower bound of 6000 cfs had approximately 1000 kg/min of sediment moving through it.  The distribution of discharge rates by clast size and discharge rate can be seen in Figure 3.
Figure 3 - Histogram of sediment discharge rate per grain size, an output of BAGS. Blue bars respresent sediment discharge rates for the calculated upper bound of 16000 cfs; Red bars represent sediment discharge rate for calculated lower bound of 6000 cfs.
            The estimated volume of the mouth bar at the time of the survey was calculated to be 1190 m­2 by,

Assuming an average density associated with sedimentary rocks of 2500 kg/m3, we can calculate time of deposition. At the calculated upper bound sediment transport rate of 6500 kg/min, the bar would take about 7.5 hours to accumulate the estimated volume.  At the lower bound sediment transport rate of 1000 kg/min, the bar would take about 45 hours to accumulate.

Figure 4 - Three dimensional surface plot of the Lester River mouth bar, assuming linear relationships. Measurements are taken with a stadia rod and eye level.
            Obtained cross-sections and photographic documentation contain many indicators of the flood suggesting a flood height of 4.2 meters.  Large point bars, large woody debris, and half-meter boulders were found throughout the study area.  
The distribution of sediment transportation rates by grain size followed the cumulative percent curve. Meaning the highest rate of sediment flux occurred with the D50 value of 64mm in both the upper and lower bound scenarios.  
In attempting to understand where the actual river discharge was in relation to our upper and lower bounds we analyzed the cross section produced by each, particularly the maximum flow depth output from BAGS which was 3.2 m at 16,000 cfs and 2.0m at 6000 cfs.  A high flood marker was found to be 4.2 meters flood water depth which points towards the upper bound being a potentially more accurate average discharge rate.  It is likely that the actual discharge was even larger than the 16,000 cfs. Using BAGS to find a corresponding flow rate to match the 4.2 meter flood height, a discharge of 18,000 cfs was found, moving 8100 kg of sediment per minute. 
            The 7.5 hour time period that it would take to deposit the measured mouth bar at the upper discharge boundary entered into BAGS is a reasonable time interval.  The river was not truly at this exact discharge rate steadily for eight hours and was likely fluctuating, but it is a good average rate. 7.5 hours is a reasonable amount of time when compared with the hydrograph of similar rivers in the area.  Lower bound inputs produced much lower sediment transport rates which were insufficient for creating a mouth bar of the observed size in a reasonable amount of time.  Therefore, the higher bound is again indicated to be closer to reality.  

            Lester River during the June 2012 flood was capable of moving clasts ranging from 8 mm to half a meter, accounting for a calculated sediment flux of 6500-8100 kg/min. Observed high flood markers from field and remote measurements suggest the max flood stage height was about 4 meters, ranging between 3.2 and 4.2 meters during peak flow. Based on our estimate volume of the mouth bar, these data suggest the material was deposited in 6-7.5 hours. We believe this number to be accurate for this time periods fits well with the hydrograph of the Knife River during this event, a similar Duluth bedrock stream.

Chow, V.T., 1959 Open-channel hydraulics: New York, McGraw- Hill Book Co., 680 p.
Czuba, Christina R., James, D. Fallon, and Erich W. Kessler. "Floods of June 2012 in Northeastern Minnesota."SIR 2012-5283. U.S. Geological Survey, 28 Dec 2012. Web. 17 Dec 2013. <http://pubs.usgs.gov/sir/2012/5283/sir2012-5283.pdf>.
"Duluth June Weather, 2012." Duluth June Weather, 2012 AccuWeather Forecast for MN 55802. AccuWeather.com, 17 Dec 2013. Web. 17 Dec 2013. <http://www.accuweather.com/en/us/duluth-mn/55802/june-weather/329420?monyr=6/1/2012&view=table>.
"Flooding Rains in Northeast Minnesota: June 19-20, 2012." Minnesota Climatology Working Group. University of Minnesota, 17 Dec 2013. Web. 17 Dec 2013. <http://climate.umn.edu/doc/journal/duluth_flooding_120620.htm>.
Graning, Amanda, Hluchan, Rick. "Summary of Duluth Flash Flood Event." June 2012 Flood in Duluth and Northland. www.crh.noaa.gov, 10 Jul 2012. Web. 17 Dec 2013. <http://www.crh.noaa.gov/images/dlh/StormSummaries/2012/June19_flood/DuluthSummary.pdf>.
Pitlick, John; Cui, Yantao; Wilcock, Peter. 2009. Manual for computing bed load transport using BAGS (Bedload Assessment for Gravel-bed Streams) Software. Gen. Tech. Rep. RMRS-GTR-223. Fort Collins, CO: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station. 45 p.
Wolman, M.G., 1954. A Method of Sampling Coarse River-Bed Material. Transactions of the American Geophysical Union 35(6):951-956