Introduction
Built in the South Fork Zumbro River floodplain in southeastern Minnesota, the City of
Rochester has experienced more than 20 major floods throughout the course of its history (Lisk
2000). Over two days in early July of 1978, intense rains fell on Rochester, including a three
hour period in which 5 inches were recorded at the airport (USGS 2008). The resulting floods
drove 5000 residents from their homes, killed 5 people, and caused $79 million in damage (List
2000 and Buckley 2004). Today, two-thirds of the city is located in the river floodplain (Buckley
2004).
In response to the 1978 flood, Rochester, Olmsted County, the Minnesota Department of
Agriculture, the US Natural Resource Conservation Service (NRCS) and Army Corp of
Engineers (USACE) combined forces to design and implement the South Zumbro Watershed
Flood Control Project. The project consists largely of upstream reservoirs designed by the
NRCS and channel improvements made by USACE. Approved for operation in 1982 [by US
congress?] (probably by the SZWFCP board), the project was completed in 1996 at a cost of
$124 million (Rainford, 2004). The project is credited with saving the city millions of dollars in
additional flood damage and won an Outstanding Civil Engineering Achievement Award of
Merit from the American Society of Civil Engineers in 2000 (Hunt 2000).
FIGURE 1 The 1978 flood near its peak (City of Rochester).
QuickTime™ and a
TIFF (Uncompressed) decompressor
are needed to see this picture.
Brief history of the political and administrative aspects of the project
The NCRS portion of the project cost came to $26 million, which includes design, construction,
securing land rights, and implementing a sediment reduction program (Ulman 2008). The
USACE part of the project cost $100 million (Rainford 2004). The City of Rochester and
Olmsted County are responsible for the maintenance and operation costs of the project. The city
has a one half cent sales tax, which, among other services, funds maintenance on the dams.
Because of the flood control project, the 2007 floods that wreaked havoc on much of
Southeastern Minnesota did little damage in Rochester. An estimated $70 million in damages
was avoided (Ulman 2008).
The U.S. Watershed Protection and Flood Prevention Act (Public Law 83-566) allows the NRCS
to provide technical and financial assistance to local governments seeking to implement flood
protection programs (NRCS 2005). With this authority, NRCS (known at the time as the Soil
Conservation Service) designed and built 7 dams in the South Zumbro Watershed: 6 for flood
control and one that provided flood control and recreational opportunities (Buckley 2004).
The dams are located on 4 of the tributaries of the South Fork Zumbro River, all of which
converge with the river in the City of Rochester: Cascade Creek, Silver Creek, Bear Creek,
Willow Creek (SCS 1990). The dams reduce the peak flows of the Zumbro River by 16%,
Cascade Creek by 21%, and Bear Creek by 35% (Ulman 2008 and Lisk 2000).
NCRS designed the dams to hold 100 years of sediment before they needed to be cleaned out
(Copeland 2008). To ensure that the reservoirs did not become clogged, the local NCRS office
instructed landowners in how to reduce the amount of erosion coming off of their lands. By
encouraging little- and no-till practices and strip cropping, 50% of the watershed is now at the
tolerable level of 5 tons/acre-year of soil erosion (Copeland 2008).
Project Summary
One of the three flood control dams on Cascade Creek, known as KR-3, is the subject of our
project and this report. The purpose of the project was to familiarize ourselves with the
concepts of watershed engineering. To do so we investigated the existing KR3 structure and its
watershed, and then determined for ourselves some of the important parameters needed to design
KR3…
Design Info: If there were two 100-year rainfall events spaced 10 days apart, the water level in
the KR3 reservoir would be at the bottom of the spillway. The top of the dam was designed for
the probable maximum precipitation: 26” in one day.
Watershed
200,540 acres (Buckley 2004)
topography: varies from nearly level to steep (SCS brocure),
geology: “much of the drainage area [of entire Zumbro watershed] is within a geologic area
known as the “Driftless Area,’ with topography comprised of a unique landform known as
‘krast.’ Krast features are characterized by numerous underground streams, sinkholes, blind
valleys and springs, and are highly susceptible to groundwater contamination”
soil types: mostly well drained silt loam
land use: largely agricultural, pasture…
climate: greater Zumbro watershed receives between 29 and 33 inches per year. (NRCS Zumbro
Watershed)
AWARDS & ACCOLADES
Minnesota Watershed Recognized By American Society of Civil Engineers - The South
Zumbro Watershed Project in southeastern Minnesota received an Outstanding Civil Engineering
Achievement Award of Merit from the American Society of Civil Engineers.
The project, located in the city of Rochester, was one of five honored during an award ceremony
in Washington D.C. on April 29. Several representatives from the Federal, State and local
partnership that worked on the project attended the ceremony. Minnesota representatives
included Rochester Mayor Charles Canfield; City Councilmen Dennis Hanson and Dave Semjen;
NRCS State Conservationist William Hunt; Robert Romocki, NRCS Engineer, Southeast
Minnesota; and two representatives from the Army Corps of Engineers. The strong partnership
exhibited during the project was praised during the ceremony. Partners included Army Corps of
Engineers, Olmsted and Upper Zumbro Soil and Water Conservation Districts, Minnesota
Department of Agriculture, City of Rochester, and Olmsted County.
The South Zumbro Watershed Project in Dodge and Olmsted counties consists of seven floodcontrol structures placed strategically along streams in the watershed. NRCS contributed $24
million toward the project; the Corps of Engineers also completed $100 million in channel work
in the watershed. South Zumbro, completed in 1996, provides flood protection, flood prevention
and recreation to residents in Dodge and Olmsted counties.
Watershed Response
Solution approach
The overall goal of this project is to reduce the downstream peak flow in the most effective and efficient
way. A reduction of peak outflow to 100 cfs for the 6.17 inch (100 year, 24 hour) storm was deemed
necessary.
Assessing the incoming water volume was the first step in the design process. After delineating our
watershed using a topographic map, we split the total watershed into five sub-watersheds using six
given structure locations, shown in Figure 2.
Figure 2: Sub-watershed Specifications
Parameters
Several parameters are required to perform watershed analysis: the area of the watershed, the time of
concentration (tc), and the watershed curve number (CN). We found the areas of the sub-watersheds
using a digitizer. Also, using a topographic map, we determined the time of concentration for each subwatershed. The curve numbers were then calculated using the SCS tables with both observed and
mapped land-use/land-cover (LULC) information.
Analysis
For our design, we considered three options: a single structure design located at structure 6 (KR-3), a
two-structure design including the upstream structures 1 and 3, and a multiple-structure design
including structure 6, structure 1, and structure 3. Upon SEDCAD analysis of these systems, we
determined that a single structure design is the most effective method of flood control.
Table 1, below, summarizes the peak discharge, runoff depth, and runoff volume for each subwatershed as well as the cumulative values (shown in bold). See the appendix for the SEDCAD outputs.
Table 1: Summary of watershed runoff values – standard values
Structure
(#)
1
2
3
4
5
6
Subwatershed(s)
(#)
Peak Discharge
(cfs)
Runoff Depth
(in)
Runoff Volume
(ac-ft)
1
347.75
3.43
41.79
1
347.75
3.43
41.79
2
222.00
1.37
28.24
1&2
551.44
3.39
70.02
3
601.91
3.23
98.41
3
601.91
3.23
98.41
4
274.45
0.80
31.96
3&4
777.47
3.28
130.37
1, 2, 3, & 4
1328.91
3.31
200.39
6
735.86
3.43
88.06
All
1829.24
3.35
288.45
To account for potential errors in our calculations of curve numbers and times of concentration, we
performed two additional SEDCAD runs using inflated values of each, holding the other value constant.
Table 2, below, shows these same values but with an increase of 15% in all curve numbers. Table 3
shows these same values with a 15% increase in all times of concentration.
Table 2: Summary of watershed runoff values – CN increased by 15%
Structure
(#)
1
2
3
4
5
6
Subwatershed(s)
(#)
1
1
2
1&2
3
3
4
3&4
1, 2, 3, & 4
6
All
CN*1.15
86.00
85.00
84.00
86.00
86.00
Peak Discharge
(cfs)
458.99
458.99
295.83
732.66
815.24
815.24
360.74
1056.69
1789.35
971.19
2457.66
Runoff Depth
(in)
4.57
4.57
1.83
4.53
4.36
4.36
1.07
4.41
4.45
4.57
4.48
Runoff Volume
(ac-ft)
55.73
55.73
37.86
93.59
132.71
132.71
42.63
175.33
268.93
117.44
386.37
Runoff Depth
(in)
3.43
3.43
1.37
3.39
3.23
3.23
0.80
3.28
3.31
3.43
3.35
Runoff Volume
(ac-ft)
41.79
41.79
28.24
70.02
98.41
98.41
31.96
130.37
200.39
88.06
288.45
Table 3: Summary of watershed runoff values – tc increased by 15%
Structure
(#)
1
2
3
4
5
6
Subwatershed(s)
(#)
1
1
2
1&2
3
3
4
3&4
1, 2, 3, & 4
6
All
Tc*1.15
(hours)
0.548
0.632
0.988
0.501
0.540
Peak Discharge
(cfs)
329.42
329.42
208.42
522.88
568.20
568.20
261.25
740.97
1789.35
698.37
1685.14
From the above two tables of data it can be seen that an increase in curve number has a much greater
impact on peak discharge and runoff volume. Clearly, time of concentration has no effect on runoff
volume since it does not change the infiltration (i.e. the same excess depth occurs with a constant curve
number), while a higher curve number correlates with higher excess depth. Thus, it is much more
important to be accurate or appropriately conservative with curve numbers in a design than with time
of concentration.
One interesting observation regarding the runs involving CN and tc inflation is that the peak discharges
through structure 5 are identical, even though the individual discharges through structures 1, 2, 3, and 4
are uniformly higher for the CN increase run. This may be purely coincidental, or could have something
to do with channel routing restricting these values.
As shown in (reference appropriate outflow hydrographs in appendix), channel routing had a small
effect on peak flow. It reduced and delayed the peaks, but by a very small amount (nearly indiscernible
on the graphs) because the actual lengths of the reaches were relatively short.
Design of KR-3
Pond Storage information
To develop the elevation storage curve, NRCS data was used. This data is listed in Table 4.
Using this data, we determined the elevation storage values for KR-3. The calculations and results are
listed in the appendix (Table 1A). Similar results can be obtained by using SEDCAD, which automatically
develops an elevation-storage curve given elevation-area data. The elevation-storage curve for
structure 6 (KR-3) is shown in Figure 3.
Table 4. Elevation-area values for KR-3
Elevation
(feet)
1092
1096
1100
1104
1108
1112
1116
1120
Surface Area
(acre)
0.14
2.39
11.39
23.65
36.50
51.93
66.12
85.12
Figure 3. Elevation Storage Curve for Structure 6 (KR-3).
Structure Six (KR-3) Elevation-Storage Curve
1000
900
800
Storage (ac-ft)
700
600
500
400
300
200
100
0
1092
1097
1102
1107
1112
1117
Elevation (ft)
To determine the volume of the sediment, the NRCS Sediment Yield Curve, which is included in
the appendix (Figure 1A), was used. Land use curve three and a reservoir sediment trap efficiency of
98% were used, and the total sediment mass and volume were computed for a deposition period of 100
years. It was also assumed that 80% of the sediment is submerged and had a bulk density of 50 lb/ft3
and that 20% of the sediment was aerated and had a bulk density of 75 lb/ft3. With these assumptions
it was approximated that the bulk density was 60 lb/ft3. These calculations are listed in the appendix
(Table 4A). The calculated sediment storage for KR-3 was 117.71 ac-ft.
Spillway Design
The elevation corresponding to the crest of the principal spillway can be determined using the
elevation-storage curve and the calculated sediment volume, by setting the spillway crest elevation at
the same elevation as the stored sediment. The value found using this method was 1104.7 feet.
However, in our design the NRCS approximated value of 1108.5 feet was used, upon the
recommendation of an executive in our engineering firm. Use of this higher elevation results in
increased permanent pool depth and a larger dam.
The diameter of the principal spillway was designed by specifying a target peak outflow of 100
cfs in response to the 6.17 inch (100 year, 24 hour) rainfall event. By trial and error using standard
concrete (Manning’s n=0.011) pipe diameters in SEDCAD, the riser diameter of 48 inches and the barrel
diameter of 30 inches were selected.
The peak water surface elevation from the 100 year, 24 hour rainfall event corresponds to the
elevation of the emergency spillway crest. This elevation was determined to be 1112.5 feet by the
SEDCAD simulation.
Design of the emergency spillway was based on NRCS guidelines. The design is based on a 6
hour rainfall depth for a class C structure. For this location, the rainfall depth was calculated to be 9.77
inches. The calculation for this rainfall depth is in the appendix. This depth was used in a SEDCAD
simulation to determine the peak water surface elevation. This peak water elevation is used to design
the emergency spillway embankment height. This elevation was 1115.1 feet.
The elevation of the top of the dam was determined using the 6 hour probable maximum
precipitation. For this location, this rainfall depth was 23.7 inches. SEDCAD was again used to simulate
the event, and the peak water surface elevation was used to determine the elevation of the dam. This
elevation was 1120.0 feet.
Generate a schematic
Hydrologic Impact
The peak outflow from the watershed was greatly reduced when KR-3 was implemented in a
SEDCAD simulation. For the 100 year, 24 hour storm, the peak outflow rate without the structure was
1829 cfs, and the peak outflow was 103 cfs with the structure. The hydrograph is shown in Figure 4.
This reduction in peak flow will help prevent flooding in Rochester, MN and other nearby communities.
Figure 4. Outflow and Inflow at Structure 6 (KR-3).
Pond Storage Information of the Upstream Structures
To develop the elevation storage curves for the two upstream ponds, elevation-area data was
used. This data was determined by using GIS and a digitizer tablet to find areas bounded by contours on
the topographic map of the site. This data is listed in Table 5. Using this data, we determined the
elevation storage values for the upstream structures. The calculations and results are listed in the
appendix (Tables 2A and 3A). Similar results can be obtained by using SEDCAD, which automatically
develops an elevation-storage curve given elevation-area data. The elevation-storage curve for
structures 1 and 3 are shown in Figures 5 and 6.
Table 5. Elevation-area values for upstream structures 1 and 3.
S1
S3
Elevation
Area
Elevation
Area
(ft)
(ac)
(ft)
(ac)
1150
0.96
1140
0.97
1160
6.52
1150
12.12
1170
19.09
1160
37.33
1180
36.85
Figure 5. Elevation-storage curve for structure 1.
Structure 1 Elevation-Storage Curve
450
400
Storage (ac-ft)
350
300
250
200
150
100
50
0
1145
1150
1155
1160
1165
1170
1175
1180
1185
Elevation (ft)
Figure 6. Elevation-storage curve for structure 3.
Structure 3 Elevation-Storage Curve
350
300
Storage (ac-ft)
250
200
150
100
50
0
1135
1140
1145
1150
1155
1160
1165
Elevation (ft)
The sediment volume was calculated in the same manner as for KR-3 (see previous section). The
calculated sediment storage for structure 1 was 10.31 ac-ft and was 21.80 ac-ft for structure 3.
Spillway Design of the Upstream Structures
The elevation corresponding to the crests of the principal spillways were determined using the
same method as was used in the design of KR-3. The values found using this method were 1154.0 feet
and 1144.5 feet for structures 1 and 3, respectively.
The diameters of the principal spillways were designed by specifying a target peak outflow of
10% of the peak inflow rate in response to the 6.17 inch (100 year, 24 hour) rainfall event. By trial and
error using standard diameters of corrugated metal pipe (Manning’s n=0.025) in SEDCAD, riser
diameters of 36 inches and barrel diameters of 24 inches were selected for both structures.
The peak water surface elevation for each structure from the SEDCAD simulation corresponds to
the elevations of the emergency spillways. These elevations were 1153.0 and 1243.1 feet for structures
1 and 3, respectively.
Hydrologic Impact of the Upstream Structures
At structure 1, the peak flow was reduced from 348 to 34 cfs. At structure 3, the peak flow was
reduced from 602 to 48 cfs. With the upstream ponds, the peak inflow to KR-3 was reduced from 1829
to 1237 cfs, and the peak outflow at KR-3 was 99 cfs. The hydrographs for these three structures are
shown in figures 7, 8, and 9. This data shows that the upstream structures, if used alone, would provide
a 32% reduction in peak outflow from the watershed. The use of KR-3 alone results in a 94% reduction
of peak outflow from the watershed. If all three structures are used, the peak flow reduction is 95%.
From this, it is apparent that if all three structures are used, KR-3 is the structure that exhibits the
greatest control over the outflow.
The upstream structures should only be used if a decrease from 1829 to 1237 cfs is acceptable.
If further peak flow reduction to approximately 100 cfs is desired, only KR-3 should be constructed.
Under no circumstances should both the upstream structures and KR-3 be constructed, as the upstream
structures exhibit little influence on the outflow from the watershed once KR-3 is in place.
Figure 7. Hydrograph for structure 1.
Figure 8. Hydrograph for structure 3.
Figure 9. Hydrograph for KR-3 with upstream ponds in place.
Conclusion
Sources
Buckley K. L. Minnesota Flood Control Project Offers Lessons to Future Engineers. Natural
Resource Conservation Service; 2004 Oct [cited 6 Nov 2008]. Available from:
http://www.nrcs.usda.gov/NEWS/thisweek/2004/041006/mnzumbroclass.html
City of Rochester. Available from: http://www.ci.rochester.mn.us/visitors/history.asp
Copeland, Dave. Fieldtrip
Lisk, Ian. After 150 Years Minnesota City Will See End of Major Flooding Problems. 2000
June [cited 13 Nov 2008]. Water Online. Available from:
http://www.wateronline.com/article.mvc/After-150-Years-Minnesota-City-Will-See-End-o0001?
[NCRS] Natural Resource Conservation Service. Watershed Protection and Flood Prevention.
United States Dept of Agriculture; updated 2005 [cited 6 Nov 2008]. Available from:
http://www.nrcs.usda.gov/programs/watershed/
Sylvia Rainford Friday, May 26, 2000 Washington, DC. NRCS This Week
Pictures:
http://www.ci.rochester.mn.us/departments/administration/projects/flood/flood_1978_.pdf
Ulman, Elizabeth. Field trip
USDA doc on watershed
http://www.mn.nrcs.usda.gov/technical/rwa/Assessments/reports/zumbro.pdf
USDA United States Department of Agriculture. South Sumbro Watershed Flood Control
Project. 1990 Feb. USDA Soil Conservation Service St. Paul MN
USGS http://ks.water.usgs.gov/pubs/reports/wsp.2502.sum78.html
http://www.ci.rochester.mn.us/departments/administration/projects/flood/Floodcontrolinformatio
n.pdf
http://www.hq.usace.army.mil/cepa/pubs/oldpubs/Awards.htm
Appendix
Calculations and Results for Elevation-Storage Curve:
Sample calculation:
𝐴𝑖 + 𝐴𝑖−1
𝑉𝑖 =
× (𝐸𝑙𝑖 − 𝐸𝑙𝑖−1 )
2
Calculations and Results for Structure 6 (KR-3):
Table 1A. Elevation-Storage Curve calculations for Structure 6 (KR-3).
Elevation
(ft)
Area
(ac)
1092
0.14
1096
1100
Structure 6 (KR-3)
Avg
ΔEL
Area
(feet)
(acre)
1.265
4
5.06
6.89
4
27.56
2.39
5.06
11.39
32.62
1112
1116
70.08
102.7
4
120.3
36.5
223
44.215
4
176.86
59.025
4
236.1
51.93
399.86
66.12
635.96
75.62
1120
4
23.65
30.075
1108
Storage
Vol
(acre-ft)
0
17.52
1104
Vol
(acre-ft)
4
302.48
85.12
938.44
Calculations and Results for Structure 1:
Table 2A. Elevation-Storage Curve calculations for Structure 1.
Structure 1
Avg
Storage
Elevation
Area
ΔEL
Vol
Area
Vol
(ft)
(ac)
(feet)
(acre-ft)
(acre)
(acre-ft)
1150
0.96
0
3.74
1160
37.4
10
128.05
19.09
165.45
27.97
1180
37.4
6.52
12.805
1170
10
10
279.7
36.85
445.15
Calculations and Results for Structure 3:
Table 3A. Elevation-Storage Curve calculations for Structure 3.
Structure 3
Avg
Storage
Elevation
ΔEL
Vol
Area (ac)
Area
Vol
(ft)
(feet)
(acre-ft)
(acre)
(acre-ft)
1140
0.97
0
6.545
10
65.45
1150
12.12
65.45
24.725
1160
37.33
10
247.25
312.7
Calculations of sediment storage volume and mass:
𝑀𝑎𝑠𝑠 𝑑𝑒𝑝𝑜𝑠𝑖𝑡𝑒𝑑 = (𝐶𝑢𝑟𝑣𝑒 3 𝑣𝑎𝑙𝑢𝑒) × 𝑑𝑒𝑠𝑖𝑔𝑛 𝑙𝑖𝑓𝑒 × 𝐴𝑟𝑒𝑎 × 98%
𝑀𝑎𝑠𝑠 𝑑𝑒𝑝𝑜𝑠𝑖𝑡𝑒𝑑
𝑉𝑜𝑙𝑢𝑚𝑒 𝑑𝑒𝑝𝑜𝑠𝑖𝑡𝑒𝑑 =
𝑙𝑏
60𝑓𝑡
3
Figure 1A. NRCS Sediment Yield Curve.
Table 4A. Sediment storage mass and volume, and principal spillway elevation.
Area
Structure (sq mi)
S1
0.21
S3
0.52
S6 (KR-3)
1.47
Curve 3
Value
Mass
Volume
P.S.
(tons/sq Deposited Deposited Elevation
mi/y)
(tons)
(ac-ft)
(ft)
1300
13477
10.31
1154.0
1100
28488
21.80
1144.5
1050
153825
117.71
1104.5
Calculations for the 6 hour rainfall depth for an NRCS class C structure:
𝐶𝑙𝑎𝑠𝑠 𝐶 𝑑𝑒𝑝𝑡ℎ = 𝑃100 + 0.26 × (𝑃𝑀𝑃 − 𝑃100 )
9.77 𝑖𝑛𝑐ℎ𝑒𝑠 = 4.88 𝑖𝑛𝑐ℎ𝑒𝑠 + 0.26 × (23.7 𝑖𝑛𝑐ℎ𝑒𝑠 − 4.88 𝑖𝑛𝑐ℎ𝑒𝑠)