Linking high temporal resolution flux-based sediment ... · boundary eroded in all time periods...
Transcript of Linking high temporal resolution flux-based sediment ... · boundary eroded in all time periods...
Linking high temporal resolution flux-based sediment budgets with channel change: Establishing morphological meaning to measurements of sediment flux
Christy Leonard 1, David Topping 2, Ronald Griffiths2, and Jack Schmidt1
Department of Watershed Sciences, Utah State Univeristy1,
Grand Canyon Monitoring and Research Center, U.S. Geological Survey2
Christy LeonardDepartment of Watershed Sciences – Utah State University E-mail: [email protected]://qcnr.usu.edu/wats/colorado_river_studies/
1
EP33D-2437
With continuous measurements of sediment transport and an innovative method to calculate a morphological sediment budget, what more can we infer about river processes?
Area to volume using a Bayesian model: Making the most of little data
Probabilistic approach to account for uncertainty in digitizing and image co-registration
Study Area:
7.1 7.12 7.14 7.16 7.18
Easting 10 5
4.4795
4.48
4.4805
4.481
4.4815
4.482
4.4825
4.483
4.4835
4.484
4.4845
Nor
thin
g
10 6
-8
-6
-4
-2
0
2
4
6
8
x
7.1 7.12 7.14 7.16 7.18
10 5
4.48
4.481
4.482
4.483
4.484
4.485
10 6
-1
-0.5
0
0.5
1
1.5
2
2.5
y
Ɛx
Ɛy
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5
Area (m 2) 10 4
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Prop
ortio
n
Veg Island Deposition
Veg Island Erosion
Channel Boudnary Deposition
Channel Boundary Erosion -1 0 1 2 3 4 5 6 7
Detrended Elevation (m)
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
Prob
abilit
y
Bank Elevation
Channel Bed Elevation
Bank Attahced Bar Elevation
In-Channel Bar Elevation
Vegetated Island Elevation
Height(vegetated island) = z(vegetated island) - z(in-channel bar)
Height(bank) = z(bank) - z(channel bed)
Height(bank attached bar) = z(bank attached bar) - z(in-channel bar)
Bayesian Model: Gibbs sampling with 10,000 MCMC stepsz ~ normal(μ,τ) – likelihood distribution
μ ~ normal(0,1x106) – uninformative prior distribution τ ~ gamma(0.01,0.01) – uninformative prior distributionz(i) ~ normal(μ , τ)
Results: Partitioned morphological and flux-based sediment budgets
2014 2015 2016-8
-7
-6
-5
-4
-3
-2
-1
0
1
Met
ric T
ons
10 5
0
50
100
150
200
250
300
350
400
450
Dis
char
ge (m
3s
-1)
Total Load
2013 2014 2015 2016-3
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
Met
ric T
ons
10 5
0
50
100
150
200
250
300
350
400
450
500
Dis
char
ge (m
3s
-1)
V Fine Sand
Fine Sand
Med Sand
Crs Sand
2011 to 2017
-4 -3 -2 -1 0
Net Volumetric Change (Metric Tons) 10 5
0
0.005
0.01
0.015
0.02
0.025
0.03
Prob
abilit
y
Total Change
Channel Change
Vegetated Island Change
2011 to 2013
-1.5 -1 -0.5 0 0.5 1 1.5
10 6
0
0.01
0.02
0.03
0.04
2013 to 2015
Flux
Bud
get
Upp
er B
ound
s
Low
er B
ound
s
-5 -4 -3 -2 -1 0
10 5
0
0.01
0.02
0.03
0.042015 to 2017
Flux
Bud
get
Upp
er B
ound
s
Low
er B
ound
s
-4 -3 -2 -1 0
10 5
0
0.005
0.01
0.015
0.02
0.025
0.03
7.126 7.128 7.13 7.132 7.134 7.136 7.138 7.14 7.142 7.144
10 5
4.4796
4.4797
4.4798
4.4799
4.48
4.4801
4.4802
4.4803
10 6
Veg Island Deposition
Veg Island Erosion
Channel Boudnary Deposition
Channel Boundary Erosion
-1 -0.5 0 0.5 1 1.5 2 2.5 3 3.5
Predicted Elevaton for Polygoni (m)
0
0.01
0.02
0.03
0.04
0.05
0.06
Pro
babi
lity
Vegetated Island Elevation
In-Channel Bar Elevation
0.5 1 1.5 2 2.5 3 3.5
Predicted Height of Deposition for Polygon i (m)
0
0.01
0.02
0.03
0.04
0.05
0.06
Pro
babi
lity
0.5 1 1.5 2 2.5 3 3.5
Predicted Volume of Deposition for Polygoni (m
3)
0
0.01
0.02
0.03
0.04
0.05
0.06
Pro
babi
lity
0 5000 10000 15000 20000
Predicted Volume of Deposition for All Polygons (m3
)
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
Prob
abilit
y
• Uncertainty was characterized in two components: (1) image co-registration error (Figure 2 A-C) and (2) digitizing error (Figure 2D).
• A spatially continuous error surface of image co-registration error (εx and εy) was created using each tie point’s co-registration error in the X (εx) and Y (εy) direction (Figure 2B).
• The channel boundary was moved by the resultant εx + εy vector (Figure 2C).• Digitizing error was characterized for each vertex by randomly sampling a X and Y
coordinate from a normal distribution over 100 iterations (Figure 2D). • 100 realizations of channel boundaries were used to compute a distribution of
planform changes between all time periods (Figure 2E).
A.
B.
100 Pr(Δarea)
X Y
C.D.
E.
Figure 3: Elevations of different geomorphic units extracted from the 2015 detrended LiDAR dataset (Figure 1). In-channel bed elevations were derived using optical bathymetry. Note: the distribution of elevations between 2011 to 2015 were similar, indicating little vertical accretion.
• A Bayesian model was used to predict the depth of erosion or deposition at each polygon for the 100 realization of channel boundaries derived in Figure 2E.
• The final posterior distribution was a combination of all realizations of channel boundaries in Figure 2E and incorporated uncertainty in image co-registration, digitizing, and depth of erosion or deposition (Figure 5).
Figure 6: (A) The morphological sediment budget showed that vegetated islands expanded and the channel boundary eroded in all time periods except 2011 to 2013 . (B) The total load flux-based sand budget was negative for all years, indicating that sand sediment was evacuated from the reach. The partitioned sand budget indicated fine and medium sand grain-size fractions were mostly in a deficit and very fine sand accumulated during high flows in 2014 and 2016.
Does change occur uniformly, or are there hotspots of change?
0.00
2
0.003
9
0.007
8 0.
016
0.03
1 0.
063
0.13
0.25 0
.5
1
2
Grain Size (mm)
0
10
20
30
40
50
60
70
80
90
100
% F
iner
Silt
VFine Sand
Fine Sand
Medium Sand
Coarse Sand
VCoarse Sand
Bank
Bank Attached Bar
Vegetated Island
In-channel Bar
Bed Material
-2
-1
0
1
2
Met
ric T
ons
10 4 2011 to 2013
0.2
0.3
0.4
0.6
0.7
0.9
1 1.2
1.3
1.5
1.6
1.8
1.9
2.1
2.2
2.4
2.5
2.7
2.8
3 3.1
3.3
3.4
3.6
3.7
3.9
4 4.2
4.3
4.5
4.6
4.8
4.9
5.1
5.2
5.4
5.5
5.7
5.8
6 6.1
6.3
6.4
6.6
6.7
6.9
7 7.2
7.3
7.5
7.6
7.8
7.9
8.1
8.2
8.4
8.5
8.7
8.8
9 9.1
9.3
-1.5
-1
-0.5
0
0.5
1
1.5
Met
ric T
ons
10 4 2013 to 2015
0.2
0.3
0.4
0.6
0.7
0.9
1 1.2
1.3
1.5
1.6
1.8
1.9
2.1
2.2
2.4
2.5
2.7
2.8
3 3.1
3.3
3.4
3.6
3.7
3.9
4 4.2
4.3
4.5
4.6
4.8
4.9
5.1
5.2
5.4
5.5
5.7
5.8
6 6.1
6.3
6.4
6.6
6.7
6.9
7 7.2
7.3
7.5
7.6
7.8
7.9
8.1
8.2
8.4
8.5
8.7
8.8
9 9.1
-2
-1.5
-1
-0.5
0
0.5
1
Met
ric T
ons
10 4 2015 to 2017
Distance Downstream of Little Snake Confluence (km)
0.2
0.3
0.4
0.6
0.7
0.9
1 1.2
1.3
1.5
1.6
1.8
1.9
2.1
2.2
2.4
2.5
2.7
2.8
3 3.1
3.3
3.4
3.6
3.7
3.9
4 4.2
4.3
4.5
4.6
4.8
4.9
5.1
5.2
5.4
5.5
5.7
5.8
6 6.1
6.3
6.4
6.6
6.7
6.9
7 7.2
7.3
7.5
7.6
7.8
7.9
8.1
8.2
8.4
8.5
8.7
8.8
9 9.1
9.3
Figure 7: (A) Distance downstream of the Little Snake confluence. (B) Total volumetric change in budget cells between image dates. Generally, areas of erosion were clustered in the upper portion of the reach from 2-2.5 km.
Figure 1: Map showing the study area located near Dinosaur National Monument on the Yampa and Little Snake Rivers. The study area is part of a larger network of USGS-GCMRC acoustic sediment gages that form four flux-based sediment budgets in the region.
Figure 2: Method to characterize and correct for spatially distributed image co-registration error (A,B) while probabilistically accounting for digitizing error (D). Areas of erosion and deposition were calculated by subtracting the channel boundary between two image dates. (E)
Figure 4: (A) Area of vegetated island deposition. (B) Predicted elevation of the vegetated island and in-channel bar from the Bayesian model. (C) Height of deposition. (D) Volume of deposition associated with a single vegetated island.
A. B. C. D.
Figure 5: Example of posterior distribution.
A.
B.
A.
B.
Conclusions Future Work
• Net accumulation of sediment appears to be associated with vegetated island expansion and can be linked to accumulation of very fine sand in the partitioned flux-based sediment budget (Figure 6).
• Vegetated islands are predominately composed of the finest sand grain-size fractions (Figure 9). • The channel narrowed in all river segments from 2011 to 2015 and widened from 2015 to 2017
on the Little Snake River and the Yampa River downstream of the Little Snake (Figure 10). Narrowing appears to be driven by vegetated island expansion that out paces the rate of bank erosion.
• Channel change can be linked to processes of sediment transport when the flux-based budget is partitioned to understand where and why channel change occurs.
Figure 9: Grain size distributions of geomorphic units used in the morphological sediment budget (Figures 6,7). Vegetated islands are composed of the finest sand-size fractions and bed-material is composed of the coarsest sand-size fractions.
• Collect repeat measurements of bed topography to characterize changes in bed elevation through time.
• Spatially characterize the grain size of geomorphic units to link the partitioned flux-based sediment budget with the morphological sediment budget.
• Examine the influence of changes in sediment supply grain size and bed grain size to adjustments in channel form.
2011 2013 2015 201778
79
80
81
82
83
84
85
86Little Snake River
2011 2013 2015 2017220.5
221
221.5
222
222.5
223
223.5
224
224.5
225
Wid
th (m
)
Yampa River DS Little Snake
2011 2013 2015 2017120
125
130
135
140
145Yampa River US Little Snake
Figure 10: Channel width through time. Width was calculated by subtracting the vegetated island area from the channel boundary area and dividing by the centerline length.