Modeling Bank Erosion, Mercury
Methylation/Demethylation, and Subsequent Receiving
Water Impacts
Mercury Workshop
Sponsored by DuPont
At the University of Delaware
January 4-5, 2005
J.J. Warwick & R.W.H. Carroll
Desert Research Institute, Division of Hydrologic Sciences Reno, NV
Overview
™
™
Site Description
Modeling Tools
™
™
™
™
™
™
™
™
™
RIVMOD
WASP5
MERC4
Extreme Event
Calibration
Verification
Ongoing Work
Conclusions
Recommendations
Collaborators
Tamar Barkay
Jean-Claude Bonzongo
Rosemary Carroll
Dan Crawford
Jadran Faganeli
Ken Heim
Mark Hines
Milena Horvat
Paul Lechler
Berry Lyons
Jerry Miller
Gregor Petkovsek
Rudi Rajar
David Wayne
Dusan Zagar
Funding Sources
NIEHS
™ NSF
™ USEPA
™ NSF
™ NSF
™
Carson River
Flow
™
Modeled domain from USGS gage
CCG through Lahontan Reservoir
(~110 Km)
™
Semi-arid river with peak flows
generally occurring in the spring
™
Catastrophic floods (e.g.. 1997flood) are generated with rare,
rain-on-snow events that occur
during the winter months
™
The meandering river is
entrenched with steep sides of
complexly structured alluvial fill
Flow
WASP5
RIVMOD
™ WASP5
™ MERC4
™
Model Augmentation
™
RIVMOD
™
™
WASP
™
™
™
™
Floodplain flow
Real sediment transport
capabilities (3 separate
particles and colloids)
Bank moisture history
Overbank Deposition
MERC4
™
No changes
Minimal Calibration
™
Bank Erosion
Evaluate fine sediment
and areal erosion
estimates
™ Adjust 3 parameters
™
™
Mercury
May 1994 (Med. Flow)
™ June 1995 (High Flow)
™ Adjust 2 parameters
™
1997 Flood
25,000
700
600
20,000
3
500
15,000
400
300
10,000
200
5,000
100
Year
2010
2000
1990
1980
1970
1960
1950
1940
1930
1920
1910
0
1900
0
Discharge (m3/s)
Discharge (ft /s)
Estimated 100-Year 1997 Flood
Flood
1997 Flood
Photos by Rhea Williams, USGS
Geomorphic Survey
™
Extensive survey
conducted in the spring of
1997 using geomorphic
techniques of aerial
photography (taken in
1991 and 1997) and
floodplain mapping
Geomorphic Survey
™
River divided
into 10 reaches
based on valley
slope and
floodplain width
Modifications to RIVMOD
1. Handles a more
complex shape
2. Computes
dynamic width
adjustment in
which eroded
mass updates
channel width
3. Divided channel
approach was
applied to the
momentum
equation
Validation of RIVMOD Code Modifications
6
20
18
5
16
4
12
3
10
8
2
6
4
1
2
P revio u s R IV M O D
M o d ified R IV M O D
H E C -U N E T
0
0
0
5
10
15
20
25
T im e (d ay)
30
35
40
45
Depth (m)
Depth (ft)
14
Modeling Bank Erosion: In-Channel Flows
Assumptions
™
™
The mass erosion rate, MER (Kg/s) is proportional to the shear stress
applied to the bank
MER is inversely proportional to the square-root of the channel bottom
slope
M ER =
ψ 1 ρ sγ w n 2 D 2 3 v 2 L s
S 01 / 2
Where D is the water depth starting at the vertical face of the channel bank
(m), and So is the bottom slope, v is the water velocity (m/s), n is
Manning’s coefficient, and LS is the segment length (m)
Modeling Bank Erosion: Overbank Flows
A second term was added to account for the underlying change of character
as the river exceeds bankfull flow (Ervine et al., 2000) such that, when D>h
M E ROB =
ψ 1 ρ sγ w n 2 D 2 3 v 2 L s
S 01 / 2
+
ψ 2 ρ s γ w n 2 (D − h )v 2 L s
h 1 3 S 01 / 2
Where h is the height of the vertical bank face.
Modeling Overbank Deposition
Course Suspended Load
Modified version of Walling and He (1997)
c
s
R
= κV C
c
s
c
main
−X f
⎞
⎛
⎜1 − e κ ⎟
⎝
⎠
Washload
WEPP (Foster et al., 1995)
R
w
s
=
β V sw
qf
(G
w
m ain
−
Tc
)
Tc
=
ψ 3 q f 2 S 01 .66
Calibration In-Channel Bank Erosion (Ψ1) and
Washload Overbank Deposition (Ψ3)
Washload Concentration (mg/L)
3,00 0
C a lib ra tio n o f ψ
C a lib ra tio n o f
1
ψ
2,50 0
2,00 0
3
U S G S O b s erv ed
M o d eled W a sh lo ad
1,50 0
1,00 0
5 00
0
0 .0 1
0 .1
1
10
1 00
3
D is c ha rge (m /s )
10 00
1 00 00
Calibration Overbank Bank Erosion (Ψ2)
Calibration of
ψ2
Modeled Values
87%
Observed Values
10
6
Sediment Eroded (10 x tons)
100
7.3%
1
4.9%
0.66%
0.1
0.12%
0.07%
0.01
0.04%
Obs. Max
Modeled
Total
Obs. Min
1997
1996
1995
1994
1993
1992
1991
0.001
Channel Width Increases
Width Increase per Unit Reach
Length (m/km)
40
30
20
10
0
Observed
-10
Predicted
-20
1
2
3
4
5
6
Reach
7
8
9
10
Overbank CSS Deposition
3.0
Observed
Predicted
Overbank CSS
6
(10 x Tons)
2.5
2.0
1.5
1.0
0.5
0.0
0
1
2
3
4
5
6
Reach
7
8
9
10
11
Overbank Washload Deposition
1.0
Overbank Washload
6
(10 x Tons)
Observed
0.8
Predicted
0.6
0.4
0.2
0.0
0
1
2
3
4
5
6
Reach
7
8
9 10 11
Sources of Hg to the Water Column
Mill Tailings
Bank Erosion
at
high flows
Q ~ 43 m3/s
Old Flood Plain
Old Flood Plain
Diffusion
Inorganic Hg from Bank Erosion
0.008
Reach 1
Reach 2
Reach 4
Reach 8
Reach 6
Reach 10
0.007
Reach 5
Reach 3
Reach 7
Reach 9
Bottom Slope
0.006
[ Hg in ]bank =
0.005
0.004
0.003
0.002
0.001
0
0
10
20
30
40
50
Distance from CCG (Km)
60
70
80
λ1
S 00.5
Mercury Speciation
Complexed
Complexed
Organic Mercury
Inorganic Mercury
Methylation
Washload
LogK
Washload
= 4.78
d
LogK
Soluble
d
= 1.19
CSS
Hg
2+
MeHg
d
= 0.32
LogK
d
= 0.90
Bedload
Bedload
LogK
= 3.63
Soluble
CSS
LogK
d
Demethylation
LogK
d
= 0.24
MeHg Bank Concentrations
™
Water content
™
Percent MeHg
™
Methylation rates
™
Non-linear contribution of
MeHg from bank erosion?
Inorganic Mercury & MeHg Calibration & Verification
100,000
10,000
1997 Flood
Overbank Flows
3
Flow at FCH (ft /s)
Bank erosion
dominates Hg
loading
1,000
100
Diffusion from
bottom
sediments
dominates Hg
loading
10
Calibration
Verification
Oct-99
Oct-98
Oct-97
Oct-96
Oct-95
Oct-94
Oct-93
Oct-92
Oct-91
1
Calibration Hgin Transport: Pre-1997 flood
λ1 = 2,500 µg/kg
100,000
May 16, 1994 (medium flow Q = 600 ft3/s)
1,000
100,000
100
10,000
10
1
0
10
20
30
40
50
60
70
Distance from Upstream Boundary (Km)
Hg (ng/L)
Hg (ng/L)
10,000
1,000
100
10
100,000
1
0
Hg (ng/L)
10,000
10
20
30
40
50
60
70
Distance from Upstream Boundary (Km)
1,000
June 10, 1995 (higher flow Q = 1,960 ft3/s)
100
10
June 16, 1994 (low flow Q = 62 ft3/s)
1
0
10
20
30
40
50
60
Distance from Upstream Boundary (Km)
70
Verification Hgin Transport: 1997 flood and beyond
100,000
Shadded region
pertain to periods of
diffusion dominated
Hg loading
10,000
1,000
8/27/97
7/8/97
5/19/97
3/30/97
2/8/97
100
12/20/96
THg at FCH (ng/L)
1997 Flood Event
Verification Hgin Transport: 1997 flood and beyond
Shadded regions pertain to
periods of diffusion
dominated Hg loading
Data associated with
Mineral Canyon flash
flood event
10,000
1,000
12/25/98
11/5/98
9/16/98
7/28/98
6/8/98
4/19/98
2/28/98
1/9/98
11/20/97
100
10/1/97
THg at FCH (ng/L)
100,000
Verification using July 23-29, 1997
(low flow Q = 43 ft3/s)
Inorganic Hg (ng/L)
10,000
1,000
100
10
1
0
10
20
30
40
50
60
70
Distance from Upstream Boundary (km)
Mercury Rates and Calibration
Methylation
Location
Water Column
-1
K20(day )
Demethylation
-1
Q10
K20(day )
Q10
Range (Km)
0
2.03
0
2.03
0 - 115.0
River Bed
0.0041
2.03
0.4483
2.03
0 - 79.25
River Bank
0.0060*
2.03
0.4483
2.03
0 - 79.25
Reservoir Bed
0.0028
2.03
1.2522
2.03
79.25 - 115.0
* calibrated
Calibration MeHg Transport: Pre-1997 flood
K20(Μbank) = 0.0060 day−1
10
MeHg (ng/L)
8
May 16, 1994 (medium flow Q = 600 ft3/s)
6
4
10
2
8
0
10
20
30
40
50
60
Hg (ng/L)
0
70
Distance from Upstream Boundary (Km)
6
4
10
2
8
MeHg (ng/L)
0
0
6
10
20
30
40
50
60
70
Distance from Upstream Boundary (Km)
4
June 10, 1995 (higher flow Q = 1,960 ft3/s)
2
0
0
10
20
30
40
50
Distance from Upstream Boundary (Km)
60
70
June 16, 1994 (low flow Q = 62 ft3/s)
Verification MeHg Transport: 1997 flood and
beyond
Shadded regions pertain
to periods of diffusion
dominated MeHg loading
20
15
10
1997 Flood
Event
5
8/27/97
7/8/97
5/19/97
3/30/97
2/8/97
0
12/20/96
TMeHg at FCH (ng/L)
25
Verification MeHg Transport: 1997 flood and
beyond
Shadded regions pertain to periods of
diffusion dominated MeHg loading
20
Data
associated
with Mineral
Canyon flash
flood event
15
10
5
12/25/98
11/5/98
9/16/98
7/28/98
6/8/98
4/19/98
2/28/98
1/9/98
11/20/97
0
10/1/97
TMeHg at FCH (ng/L)
25
Ongoing Activities
™
™
™
™
Reservoir physical
characterization
Reservoir net settling
Characterize erosion
uncertainty
Perform Monte Carlo
analysis to determine
probable range of predicted
system behavior
Conclusions
™
™
™
™
Simplistic model of bank erosion predicts in-stream sediment
concentrations well over a large flow domain and channel
widening well over a large spatial domain
Overbank deposition of Course Suspended Sediment is
predicted well without calibration over a large spatial domain
Overbank deposition of Washload is correctly over-predicted,
but this does not validate the approach
In-stream inorganic and methyl mercury concentrations are
predicted well over a large spatial and flow domain
Recommendations
™
™
Determine measures of
mercury bioavailability
Re-define mercury
methylation and
demethylation kinetics
™
™
™
First-order?
Important environmental
factors and associated
corrections
Deal with and express
impacts from parameter
uncertainties