UNIVERSITY OF MINNESOTA
ST. ANTHONY FALLS LABORATORY
Engineering, Environmental and Geophysical Fluid Dynamics
Project Report No. 400
Density Current Instrusions in an
Ice~Covered Urban Lake
by
Christopher R. Ellis, Jerry Champlin,
andiHeinz G, Stefan
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D. C.
March 1997
Minneapolis, Minnesota
TABLE OF CONTENTS
Acknowledgement................................................................................................. ii
Abstract. ...................................... ,...........................................................................iii
1.
Introduction .. ,......................................................................................................... 1
2.
Site Description....................................................................................................... 2
3.
Inflow and In-lake Salinity (Density) Measurements ............................................... 4
3.1 Study Methods and Design .............................................................................. 4
3.2 Results ....................................... ,...................................................................... 4
4.
Temperature Measurements .................................................................................. 10
4.1 Methods and Design ...................................................................................... 10
4.2 Temperature Profile Dynamics in Time .......................................................... 10
5.
Implications and Conclusions ................................................................................ 23
References ............................................................................................................. 25
1
ACKNOWLEDGMENT
The investigation described herein was conducted for the U.S. Environmental
Protection Agency as part of a research program dealing with the alteration of water
availability, water quality and fish habitat in cold regions by climate change. This particular
investigation deals with lakes in spring. Barbara M. Levinson of the USEPAlORD is the
project officer. The EPA support is gratefully acknowledged.
The University of Minnesota is committed to 111e policy 1hat all persons shaI1 have equal access to its programs, facilities, and employment without
regard to race, religion, color, sex, national origin, handicap, age or veteran status.
Prepared for
Date Submitted:
Last Revised:
Disk Locator:
USEPA .
March 10,1997
March27,1997
(a:Vl165. PR400TXT.doc; PR400CDV.doc; Tab1el.doc)
ii
ABSTRACT
Evidence is presented that snowmelt runoff from an urban watershed can produce density
current intrusions (underflows) in a lake. Several episodes of density current intrusions are
documented. Water temperatures and salinities measured near the bottom of a 10m deep
Minneapolis lake during the late winter warming periods in 1989, 1990, 1991, and 1995 show
significant rapid changes which are correlated with observed higher air temperatures and
snowmelt runoff. The snowmelt runoff entering this particular lake (Ryan Lal(e) has increased
electrical conductivity, salinity, and density. The source of the salinity is the salt spread on
urban streets in the winter. Heating of littoral waters in spring may also contribute to the
occurrence of the sinking flows, but is clearly not the only cause.
111
1. INTRODUCTION
Urban runoff can carry significant pollution loads, including suspended sediments,
nutrients and metals. The receiving water bodies are often streams and rivers of various sizes,
and in some cities urban lakes and impoundments. In these latter cases, the impact of the
polluted water input on water quality can be serious. Cities in Minnesota, for example, have
long struggled with this problem. In this geographic location a significant portion of the annual
urban runoff occurs in spring as snowmelt runoff.
In this paper we document for the first time an unexplored aspect of this snowmelt
runoff~-its appearance as a density current at the bottom of urban lakes. This has importance in
lake modeling for lake water quality management in general, since the density currents carry
materials to the bottom of a lake where they may ultimately get buried in lake sediments if the
lakes are deep and stratified, or from where they may become moved to the sutface mixed layer
if the lakes are shallow and easily mixed.
Density currents are submerged gravity driven flows which occur when inflows to a
. water body are denser than the ambient water. The inflow subsequently plunges and continues
as a distinct flow which can be envisioned as a submerged stream. Density currents, also called
underflows, are known to form intermittently on coastal continental shelves, in reservoirs and
at effluent discharge sites. They have been observed and documented in all three of these
environments. A review on density currents was recently given by Alavian et al. (1992). The
mechanics of density currents have been described in various models, e.g. by Turner (1973),
Akiyama et al. (1987) and many others. Density currents can also have an internal origin. For
example, heating or cooling of littoral waters to maximum density or resuspension of sediment
in littoral waters by waves can initiate density currents.
In a continuing investigation of temperature dynamics and transport processes in icecovered lakes, the authors have collected evidence of the existence of density currents in urban
lakes that result from snowmelt runoff. Presentation and analysis of this information is the
focus of this paper. Heating of littoral water may have been a contributing factor.
1
2. SITE DESCRIPTION
Ryan Lake, located in Minneapolis, Minnesota, illustrated in Fig. 1, has been the site of
this and previous research (Ellis and Stefan 1991). It has a surface area of 6.1 ha, a maximum
depth of 10m and a mean depth of 5 m. It is a highly eutrophic lake which occasionally
experiences winterkill; the last occurrence of winterki11 was in 1988. Ice thicknesses on the
lake reach 0.5 m maximum. As the ice-cover develops, its open water volume decreases of
305,000 decreases by up to 10%. It receives inflow from urban storm sewers (II and 12 in
Fig. 1), Intermittent Creek (13), and direct road runoff (I4). During winter and spring snow
melt, the majority of the inflow to the lake comes from one of the storm sewers (11) and the
intermittent creek from Twin Lakes (13).
nr
2
,....... ,
....
\
\.1
"'"" .........
.-
-...............
/)
I
Railroad
....... .... _--
........ ,
I
\. I
, I
/
I
'oJ
Peat Bog
I
I
I
~
!\
Storm Sewer \
Location
J
I
I
I
<:~
./
",,""
~
I
I
,
I
/
,
I
\
\
I
, •..1
I
I
--,
o
I
Figure 1
5
I
Measurement
100 meters Location
Storm
Sewer
Location
I
Bathymetry and surroundings of Ryan Lake, Minnesota. Surface water
inflow points are labeled 11 to 14. Measurement locations are H, J, and K.
3
3. INFLow AND IN-LAKE
SALINITY (DENSITY) MEASUREMENTS
3.1 Study Methods and Design
Water temperature data were obtained with a Brooklyn precISIOn mercury
thermometer, which is accurate to within ± 0.05 DC, and the conductivity measurements were
taken with an American Marine Inc. PINPOINT conductivity monitor which has a stated
accuracy of ± 5%. The correlation between conductivity and density was carried out under the
assumption that NaCl is a close approximation (in terms of added density and ionic strength) to
the salt applied to roads in Minneapolis. A Shimadzu Libror Model EB-4300DW scale was
used to detennine added density due to a range of salt concentrations. This procedure yielded
the linear relationship (Champlin 1996)
Added Density [kg/m3]
=
0.0004 ' Conductivity (mSiemens].
[1]
3.2 Results
Evidence of elevated salinity in the inflows to the lake in spring is provided in
measurements made in February and March, 1995. Temperatures, conductivities, and densities
of the inflows to the lake are shown in Table 1. High salinity and density are evident.
Maximum daytime air temperatures above ODC and nighttime air temperature minima near ODC
usually cause snowmelt runoff events when the ground is snow-covered. When such air
temperatures occur, runoffwill follow within 0 to 2 days. The air temperature record shown in
Fig. 2. provides an example of air temperature during a snowmelt period. Because the streets
in the watershed are salted in winter, the first flush runoff carries this salt with it.
If snowmelt runoff causes density currents in the lake, salinities and densities at the
bottom ofthe lake should reflect a change in salinity as time passes and density currents arrive.
Table 2 gives measurements of salinity and density near the bottom of the lake at three different
times in spring.
A detailed investigation of salinity in the vicinity of the storm sewer inflow 11 was
conducted on March 13, 1995. It provided clear evidence of a plunging phenomenon very
near the edge of the lake. Schematic drawings of this event are presented in Fig. 3, and data
collected at the locations indicated in Fig. 3 are presented in Table 3.
4
Figure 3 shows schematics in both plan view and cutaway view of the density current
observed on March 13, 1995 (Julian Day 72). The snowmelt water inflow cut through the ice
surface near the culvert. The ice cover within 30 to 40 m of the shore near the culvert was
under two to six inches of water. The majority of this overlying water is likely to be from
snowmelt on the lake. This observation is supported by the surface conductivity measurements
presented in Table 3. A density current is evident a few feet from the inflow into the lake:
conductivity measurements at 0 and 0.3 m depths at locations A, C, and D indicate that the
highly saline inflow plunges rapidly as it loses momentum entering the lake. The measurements
at Hjustity the schematic shown in Figure 3 which is based on the model by Turner (1973).
The inflow temperatures reported in Table 1 in conjunction with the observed plunging
phenomenon presented schematically in Figure 3 indicate that the probable source of the
intrusions documented in the following section is the slightly saline snowmelt runoff. Heating
of shallow littoral water through decaying ice and sediment in the surface runoff may have been
contributing factors.
5
TABLE 1. Inflow characteristics in spring
Air Temp.
max,min [DC]
Location
(Fig. 1)
Water Temp.
[OC]
Conductivity
[pSiemens]
Density
[kglm3]
2/18/95
6.1,~3.3
11
0.7
868
1000.264
3/13/95
14.4,7.8
I1
4.5
1122
1000.444
3/13/95
14.4,7.8
13
0.7
622
1000.166
3/13/95
14.4,78
14
6.4
650
1000.152
3/17/95
15.5,3.9
I1
5
220
1000.077
3/19/95
7.2,~1.1
11
1.9
80
999.9981
3/19/95
7.2,~1.1
13
4.3
565
1000.222
3/23/95
11.1,1.1
11
7.3
460
1000.097
3/23/95
11.1,1.1
13
8.3
490
1000.058
Date
6
20
10
0
-10
5' -20
0
e?
::J
ro
'Q)"'"
0-
E
~
.....
«
10
0
-10
-20
66
67
68
69
70
71
72
73
74
75
Julian Date
Figure 2
Air temperature record during the late winter 1994/95 (January 1 to March 31,
1995), including a snowmelt period from March 10 (Julian Day 69) to March 15
(Julian Day 74).
7
TABLE 2. Conductivities and densities of water near the bottom of Ryan Lake.
Air Temp.
max, min [0C]
Location
[Fig. 1]
Depth
[m]
Water
Temp·r°C]
Conductivity
[/-tSiem ens]
Density
[kg/m 3]
2/18/95
6.1 ,~3.3
K
7.8
~-
65
1000.023
2/18/95
6.1,~3.4
J
8.1
~-
67
1000.024
2/18/95
6.1,~3.5
H
6
~-
71
1000.025
2/18/95
6.1,~3.6
H
6.5
~~
63
1000.022
2/24/95
2.8,-7.8
K
7.8
78
1000.028
2/24/95
2.8,~7.9
J
8
71
1000.025
2/24/95
2.8,-7.10
H
6
-~
60
1000.021
3/12/95
15,7.8
K
8
3.1
770
1000.369
3/12/95
15,7.9
J
6.5
--
943
1000.305
Date
-~
-"
TABLE 3. Temperature and conductivity measurements on
March 13, 1995 to accompany figure 3.
Location
Depth
[m]
Water Temp.
[0C]
Conductivity
[/-tSiemens]
Density
[kg/m3]
It
0
4.5
1122
1000.4437
A
0
4.7
1000
1000.3928
B
0
1.5
232
1000.0451
C
0
1.5
242
1000.0491
D
0
1.5
236
1000.0467
E
0
1.2
208
1000.0355
F
0
1.3
200
1000.0204
G
0
1.2
200
1000.0245
H
0
~-
150
1000.0003
B
0.3
1000
1000.3973
C
0.3
---
840
1000.3334
D
0.3
--
800
1000.3174
H
0.3
612
1000.2358
8
(a) Plan view
culvert Inflow, 11
Ice
(b) Cutaway of Center Portion Illustrating Density Flow
Culvert .
H
I
T
-1m
I
Figure 3
1000.35 to 1000.45 kg/m3
1000.25 to 1000.35 kg/m 3
1000.10 to 1000.25 kg/m 3
1000.00 to 1000.10 kg/m 3
t------20m----l
1
Schematic of density current intrusion. Differences of shading indicate
variable water density.
9
4. TEMPERATURE MEASUREMENTS
4.1. Methods and Design
On February 1, 1991, an array of 20 thennistor~type temperature probes attached to a
5 m long rigid pole was positioned in and above the sediment of Ryan Lake at its deepest point
such that 2 m of the pole was buried in the sediment and 3 m extended above it. Enough
flotation was added to the pole to make it neutrally buoyant and prevent it from sinking further
into the sediment after its initial placement. With the pole properly positioned, one probe was
located at the sediment sUlface, 8 probes were located below it, and 11 probes were located
above it. Both in the sediment and water, probes were placed closer together near the
sediment/water interface where the greatest gradients in temperature were anticipated. The
error associated with probe placement was ± 1.0 em (absolute).
The thennistor probes were attached to 2 Campbell Scientific CRI0 dataloggers which
were programmed to make temperature measurements every minute. Every 20 minutes, the
last 20 measurements were averaged and stored for later retrieval. To capture perturbations in
the temperature structure, all of the 1 minute measurements made in the water column and at
the sediment surface were stored if any measured water temperature differed by more than
O.I°C from the probe's previous measurement until either the data were retrieved or 24 hours
had elapsed, whichever happened first. This measurement scheme provided both data with
relatively low resolution in time (20 minute averaged data) during periods exhibiting relatively
steady temperatures and higher temporal resolution during more dynamic periods (i.e., during
the occurrence of transients). Such a scheme was necessitated by memory limitations of the
dataloggers. Errors associated with differential measurements (successive measurements made
by the same probe) were ± 0.1 °C while absolute accuracy ofthe calibrated probes was ± O. 5°C.
4.2. Temperature Profile Dynamics in Time
A measurement record of the 1 minute data of 3 probes in the water column is shown
in Fig. 4. The quasi~steady water and sediment temperatures prior to February 4 and the
perturbation events that occurred thereafter are noteworthy. No appreciable thaw had
occurred between early December and February 4, at which time roughly 20 em of snow cover
was lost in a three~day period (Fig. 5). The first perturbation appeared in the water
temperatures after 2000 hours on 3 February and was restricted to the bottom 2 m of the water
column (Fig. 6). The record is typical of the "head" of a density current (underflow) which
begins as a very thin layer at the sediment surface and progressively thickens to 2 ill, 5 hours
10
later. A much larger perturbation was recorded commencing at 1400 hours on 4 February.
This event affected all monitored portions of the water column (0 to 3 m above the sediments)
almost immediately and dropped overall water temperatures in that layer by about 1DC.
Temperature fluctuations associated with the arrival of the intrusion damped out somewhat in
the following 23 hours when a smaller secondary perturbation appeared at 1300 hours on
February 5.
That temperature transients show the presence of density underflows is supported by
the fact that the bulk of the previously deposited snow cover melted on February 3 and 4. A
substantial inflow to the lake from the storm water culvert on the east shore was observed on
February 3. Without dissolved or suspended contaminants, this water of unknown but
probably near~freezing temperature would not be expected to plunge as a density current to the
lake bottom since O°C water is usually the lightest water under ice. However, since salt had
been applied to the streets in the watershed for the previous two months and there had been no
significant thaw to remove it, the inflow was saline causing it to be more dense, even the 4°C
water near the lake bottom.
The interpretation of the temperature record as a density current intrusion facilitated by
a plot of isotherms over depth and time (Fig. 6). Isotherms from 2.9°C - 4.4°C are plotted
covering a height of3 m above the sediment/water interface from February 2 to February 9. If
the time (horizontal) scale is interpreted as distance seen by a traveling observer or recorder,
Fig. 6 can be interpreted as a representation of the head of a density current moving from right
to left. Its shape more clearly shown in a report by Ellis et al. (1997) is similar to that reported
in the literature (Ippen, 1966; Harleman, 1960). Wave-like disturbances are readily apparent in
the shear layer where temperature gradients are the strongest.
A number of other perturbations/intrusions appear in the temperature record later in
February and in early March 1991. Again, these episodes occurred at times when air
temperatures were above freezing and snow cover was melting.
Intrusions of density currents resulting from snowmelt were observed also in the late
winters of 1989 and 1990 in which similar water temperature measurements were made. Three
events are illustrated in Figs. 7 to 12 along with their associated air temperature records.
,4
'
Air temperature records before, during and after the density current intrusion observed
in 1989 are given in Fig. 7. Air temperatures were rising during the two days preceding the
density current observation from - 10°C to + 7°C during the day. These temperatures are
sufficient to produce snowmelt runoff from city streets. Lake water temperatures had been
monitored using a thermistor chain since 8 February from the surface to the 8 m depth.
Isotherms interpolated from the data are shown in Fig. 8. The arrival of the density current on
March 10 (Julian day 69) is evidenced by a sudden water temperature drop from 3°C to 2.4°C
at the 8m depth. Propagation of the disturbance associated with the arrival of the density
current can be seen to extend all the way to the ice cover.
11
The sudden appearance of 2.4 DC water near the lake bottom underneath 2. goC water
can only be explained as an impurity laden (saline) density intrusion.
Two similar events were observed in 1990, the first on 21~22 February (Julian day 52~
53 in Fig. 10) and the second on 7 March (Julian day 66 in Fig. 12). Again, both events
corresponded with thawing (above freezing) conditions of the ambient air (Figs. 9 and 11). On
21 February, the sudden appearance of3.3°C water near the lake bottom where hours earlier
quiescent, stably stratified 4°C to 4.2°C water had existed, points to the intrusion of an
impurity~laden cold underflow originating from snow~me1t runoff. Likewise on 7 March, a
near~bottom disturbance can be seen which lags by about 24 hours behind the onset of above
freezing air temperatures. It may be noted that the uncharacteristically high water temperatures
in the upper part of the water column and the near complete lack of stratification during the
week of5~11 March (Fig. 12) were due to solar heating of the water column through a snow~
free ice cover.
12
2.5
r--I
U
o
I--J
3.0
Q)
I-
:;
+-l
2
3.5
3.0m -
Q)
........~"""""'-""~~"<"'I'jll
D-
E
~
4.0
100m
4.5
O.2m
Feb 3 ; 200
Feb 4;
0
Feb 4; 200
Feb 5;
0
Feb 5; 1200
Feb 6
00
Date;Hour-Minute
Figure 4
One minute water temperature data recorded at selected depths above the lake
bottom, 3~6 February 1991. Major cold water intrusion on 3 and 4 February.
13
20,------------------------------------------r~
-J, Appearance of Denelty C~rrent
10
Q)
L
:J
+-'
o
L
Q)
D-
-10
E
Q)
I-
-20
L
-30
1990
-.-
20
1991
Snow on Ground, Metro Area
t::::2l Daily Snowfall, Metro Area
-J,Appearance of Deneity Current
15
....c:
+-'
D-
~ 10
~
o
c
(/)
5
O~~mttmw~mw~~~mm~mw~~~~~~~~ffiTI~~~
Dec 1
Figure 5
Dec 15
Dec 29
Jan 12
Jan 26
Feb 9
Feb 23
Mar 9
Mar 23
Records of air temperatures, snowfall and cumulative snow depths during winter
1990/91 (1 December 1990 - 31 March.J991).
14
I-'
Vl
Figure 6a
Isotherms of 1 minutes water temperature measurements, illustrating on the left the front of a cold water
intrusion beginning on 3 February (Julian Day 34) 1991, followed on the right by a second intrusion on 4
February (Julian Day 35).
o
2
(No Data)
........ 4
g
:5
0.
0>
0
6
8
1---- . . _____.
10~~~~~~~~~~·~~~~
33
37
38
. 39
40
Julian Day 1991
Figure 6b
Isotherms in the bottommost 3m of water in Ryan Lake from 2 February to 9
February 1991. Stable stratification on the first two days is disturbed on the
following two days. Water is again stratified at the end of the period, but colder.
16
20
15
r-...
0
0
"--'"
Q)
I-
~
10
5
0
+'
0
IQ)
-5
0..
-10
E
Q)
I-
-15
I-
«
-20
-25
-30
B Feb
15 Feb
22 Feb
1 Mar
B Mar
15 Mar
22 Mar
29 Mar
5 Apr 1989
15
10
r-...
0
0
"--'"
Q)
I~
5
+'
0
IQ)
0
0..
E
Q)
l-
-5
I-
«
-10
..-15
Mar 7
Figure 7
Mar 12
Mar 13
1989
Record of air temperatures during'the late winter 1988/89 and at the time of the
observed cold water intrusion in Ryan Lake, 10 March 1989.
17
0
. -'\
2
~,~
2
2,8
, ,.
2~~' ~.
..-.. 4
E
'-"
.c
.....
Q)
Cl
~
~~!l2'
. 0.
6
\
3
2.4 2.4.
8
(No Data)
10
66
67
68
70
69
71
72
73
Julian Day 1989
0
Temp (0)
.~
3,1
'3
2
"......,
E
.........
2.9
2.8
2.7
2.6
2.5
2.4
2.3
2.2
2.1
2
1.9
1.8
1.7
1.6
4
..c
+-'
0..
Q)
Cl
6
8
(No Data)
10
66
67
68
70
69
71
72
73
Julian Day 1989
Figure 8
Isothenns in the upper 8m of water in Ryan Lake from 7 March to 13 March 1989. Disturbance
of stable stratification by cold water intrusion on 10 March (Julian Day 69). Stratification at end
of period shows lower temperature near bottom and thinner thennocline layer below the ice,
presumably due to mixing induced by cold water intrusion.
18
20
'I
.
'I
'I
'I
I
I
'I
1
'I
'I
'I
1
15 -
10 5\,
(J)
I;:J
+-'
o
I(J)
D-
~
I~
-5 -
,~
N
III.
o
~W~
~
1'1
I~
:.
Ir
-
E
-10 -
I-
-15,-
-
-201-
-
-25,..:
-
(J)
I-
:<c
-30
.
1990 •
Jan 1
'I
'I
Jan 15
'I
'I
Jan 29
1
'I
Feb ·12
I
1
Feb 26
'I
'I
Mar 12
'I
'I
Mar 26
10
5
,--.....
0
0
0
'-'"
Q)
I-
::J
-5
+-'
0
IQ)
-10
D-
E
-15
Q)
lI-
«
-20
-25
-30
Feb
1~
Feb 19
Feb 20
Feb 21
Feb 24 1990
)
Figure 9
Record of air temperatures during the late winter 1989/90 and at the time of the
cold water intrusion recorded in Ryan. L8.ke on 21 and 22 February 1990.
19
Or---------------------------------
------~------~
2
r
E
........
4
.c
.....
DID
o
"6
I
3,5
81=-_ _ _ _ _ __
1----~-3,8-----_il11
'_"
50
51
52
53
54
55
56
Julian Day 1990
0
Temp'(C)
4.4
4,3
4.2
4.1
2
4
--
E
..........
3.9
3.8
3.7
3.6
3.5
3.4
3.3
3.2
3.1
3
2.9
4
J::
0..
Q)
Cl
6
8
10~~~~~~~~~~
~~~
49
II
50
51
52
53
54
55
56
Julian Day 1990
Figure 10
Isothenns in the deepest water column of Ryan Lake from 18 Februar
y (Julian Day 49) to 24
February (Julian Day 55) 1990. Disturbance of the stable stratification
by a cold water intrusion
on the bottom of the lake on 21 February (Julian Day 52).
20
15~----~----~----~----~-----r-----'--~-.
,.---....
10
U
0
'--../
Q)
5
L
::J
+-'
0
L
Q)
0
Q...
E
Q)
I-
-5
L
<C
-10
-15~----~----~----~----~-----r----~----~
Mar 5
Mar 6
Mar 10
Mar 9
Mar 11
1990
Mar 7
Mar 8
{!
Figure 11
Record of air temperatures at the time of the cold water intrusion recorded
in Ryan Lake on 7 March 1990.
21
2
.-. 4
.s
-o
.t:
C.
m
6
66
65
67
69
68
70
71
Julian Day 1990
o
Temp (C)
4.2
4.1
4
3.9
3.8
2
3.7
3.6
3.5
3.4
3.3
3.2
3.1
3
2.9
........ 4
E
.........
..c:
15.
Q)
Cl
6
2.8
2.7
8
10~~~~~~~~~~
64
65
66
68
67
69
70
71
Julian Day 1990
Figure 12
Isotherms in the deepest water column of Ryan Lake from 5 March (Julian Day 64) to 11 March
(Julian Day 70) 1990. Disturbance of the stable stratification by an intrusion on the bottom of the
lake on 7 March (Julian Day 66).
22
5. IMPLICATIONS AND CONCLUSIONS
Substantial evidence has been presented that early spring warming in a cold region
urban watershed brings density current intrusion in lakes. These currents have been linked to
snowmelt runoff and increased salinity in the runoff water. The study reported herein was·
conducted in an urban area where roadsalt is regularly applied.
The lag time between above freezing air temperatures and the appearance of density
current intrusions has been observed to be 0 to 2 days. Thicknesses of these currents have
been observed to be on the order of several meters. After their passage, the entire lake
stratification is disturbed and bottom temperatures are often cooler.
The four episodes of urban snowmelt runoff density current intrusions in a Minneapolis
lake reported, occurred between February 4 and March 10. A typical ice-out date for the lake
is April 1. Spring overturn (4DC uniform lake temperature) occurs shortly thereafter.
Snowmelt runoff from urban watersheds into lake therefore precedes spring overturn
by several weeks. The reason is that a much larger portion of the incoming solar radiation in
spring is captured by the street surfaces, roofs, and building surfaces, than by the snow and icecovered lake surfaces. Often streets and buildings are free of snow and ice, while urban lakes
still have a complete ice cover. The physical reason is the difference in the albedo of street and
building surfaces (0.05 to 0.20) and snow (0.7 to 0.9) and ice (0.3 to 0.80).
It was shown herein that urban snowmelt runoff flows to the deepest portion of a lake.
Because of the timing of snowmelt runoff, it will still be at the lake bottom when the lake turns
over in spring. Being slightly more saline, the layer of snowmelt runoff water has the potential
to prevent the complete overturn (mixing) of the lake to its very bottom. Instead, the slightly
saline snowmelt water layer may cause at least a temporary mesomixis.
There is at least one lake (Brownie Lake) in the Minneapolis chain of lakes with
documented mesomixis due to street salt runoff. Brownie Lake has a year-round saline bottom
layer and is a very small and relatively deep lake set in a deep topographic depression. It is a
very wind-sheltered lake. The half-life of the snowmelt runoff layer in Ryan Lake is not known
at tllls time. In this study, we have documented previously uncaptured events which indicate
the formation of the layer. Its ultimate fate will be explored through further field studies and
modeling ofthe vertical mixing near the bottom oflakes after spring (snowmelt runotI).
23
The study reported herein has also established that in one-dimensional models of urban
lake water quality, urban snowmelt runoff water volume and its solids content can be placed
into the bottom-most layer. The likelihood that suspended materials, which snowmelt runoff
imports to an urban lake, settle out and are buried in lake sediments is therefore very high. The
distribution of dissolved materials,' e.g. nutrients from the snowmelt runoff in the lake, will
depend whether the receiving lake bottom layer resists vertical mixing or not, which can be
determined based on the vertical salinity (density) gradients, lake size and depth and weather
(wind) conditions.
'
One implication of this observation then is that in early spring (February and March) ,
while urban lakes are still mostly ice-covered, urban watershed runoff carries materials
accumulated on city streets quickly and efficiently to the bottom of lakes. In addition to
potassium and calcium chlorides, these materials may include nutrients and metals. The
concentrations are apparently small and no mesomixis during the remainder of the year has
been reported.
24
· REFERENCES
,
Akiyama,1. and H. G. Stefan. 1987, Onset of underflow in a slightly diverging channel. Jour.
Hydraulic Engineering, ASCE, Vol. 113, No, 7, 825~844.
Alavian, v., G. H, Jirka, R A. Denton, M, C. Johnson and H. G, Stefan. 1992. Density
currents entering lakes and reservoirs, Jour. Hydraulic Engineering, ASCE, Vol. 118,
No. II, 1464-1489.
Champlin, J. G. and H. G. Stefan. 1996. Field Study of the Ice Cover of a Lake: Implications
for Winter Water Quality Modeling, Project Report 387, Univ. of Minnesota, St.
Anthony Falls Laboratory, Minneapolis, MN, 212 pages.
Ellis, C. R and H. G. Stefan. 1991. Water temperature dynamics and heat transfer beneath
the ice cover ofalake. Limnol. Oceanog. 36(2),324-335.,
c
Ellis, C. R, J. G. Champlin, and H. G. Stefan. 1997. Density current intrusions in a icecovered, urban lake, Project Report 400, Univ. of Minnesota, St. Anthony Falls
Laboratory, Minneapolis, MN, 20 pages.
Harleman, D.R.F. 1960. "Stratified flow," Handbook of Fluid Dynamics, Ch. 26, Streeter, V.
L. Editor, McGraw-Hill Book Col., New York, NY., pages 26-1 to 26-21.
Ippen, A. T. 1966. Estuary and Coastline Hydrodynamics.
York, 744 pages.
McGraw~Hill
Book Co., New
Turner, J. S. 1973. Buoyancy Effects in FlUids, Cambridge University Press, 367 pages.
25