500 m
N
N
2.5 km
AirSWOT Elevation (m)126+116
AirSWOT Elevation (m)139+128Pressure Transducers
June 9 AirSWOT vs. Observations: Upstream Lines
Downstream Distance (km)
WSE (m)
010203040
126
128
130
132
134
136
138
140
June 9 AirSWOT vs. Observations: Downstream Lines
Downstream Distance (m)
WSE (m)
051015202530
116
118
120
122
124
85 80 70 65 60
124
122
120
118
116
75
West Profile 140
010203040
138
136
134
132
130
128
126
b
WSE
(m)
East Profilea
Downstream Distance (km)
GPS Profile MeasurementsAirSWOT Orthogonal MeanPressure TransducersAreas of High TopographyAirSWOT 1 SDAirSWOT 2 SD
Downstream Distance (km)
Hei
ght E
rror
: Obs
erva
tions
- A
irSW
OT
(m)
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3
-10
010
30
25457585
Downstream Distance (km)
W
SE E
rrors
(cm
)
RMSE = 9.0 cmMAE = 7.1 cm
Bias = 0.65 cm
5
20-2
0-3
0
65 55 35 15
Slope Variations - Upstream Region
Downstream Distance - Reach Center (km)
Slopes
101520253035
0.22
0.24
0.26
0.28
0.30
Downstream Distance - Reach Center (km)
Slopes
65707580
0.20
0.25
0.30
80 70 65
3025
20
75
West Profile
NSE = 0.93
ba
10152025
3028
2624
22
Downstream Distance - Reach Center (km)3035
Slo
pe (
cm/k
m)
East Profile
NSE = 0.76
GPS Profile SlopesAirSWOT Slopes
Histogram of Slope Errors along 10 km Reaches
Slope Error (m/km)
Freq
uenc
y
-0.03 -0.02 -0.01 0.00 0.01 0.02 0.03
020
4060
8010
012
012
010
080
6040
200
-3.0 -2.0 -1.0 0.0 1.0 2.0 3.0
Slope Errors (cm/km) for 10 km Reaches
Freq
uenc
y
1. Key Points • AirSWOT provides a new method for measuring river water surface elevations (WSEs) and slopes without the need for in situ data. • Errors from AirSWOT are sufficiently small to allow detection of decimeter-level variations in WSEs over 1 km areas and cm/km level variations in slopes along 10 km reaches. • Results indicate AirSWOT is capable of producing measurements useful for validating SWOT-quality measurements of river WSEs and slopes.
2. Data Analysis • Collected a GPS profile of WSEs along the main channel of the Tanana River coincident with AirSWOT measurements on June 9, 2015 (Figure 1). • Filtered AirSWOT WSEs using a ratio of radar magnitude to estimated error. This filter helps eliminate pixels effected by layover and artifacts due to the processing methodology. • Calculated orthogonal means of AirSWOT WSEs for every GPS measurement along the profile (Figure 2). • Estimated AirSWOT WSE errors when averaged over 1 km areas (Figure 3). • Used a moving window every 100 m along the GPS profile to calculate AirSWOT and GPS profile slopes along 10 km reaches, and assessed AirSWOT’s ability to capture slope variability by calculating Nash-Sutcliffe Efficiency (NSE) values between GPS Profile and AirSWOT slopes (Figure 4). • Calculated AirSWOT slope errors for the 499 overlapping 10 km reaches (Figure 5). • AirSWOT WSE and slope errors are compared to the SWOT mission’s science requirements for executing robust science of surface water dynamics. These requirements are ±10 cm or better for WSEs when averaged over 1 km areas and ±1.7 cm/km for slopes after processing along a maximum of 10 km of flow distance [Biancamaria et al., 2016; Rodriguez, 2016].
Figure 2: WSE profiles derived from GPS profile (black) and AirSWOT (red) for the (a) west and (b) east regions. Standard deviations (SD) are shown for the AirSWOT WSEs in the grey shaded areas.
Figure 4: Slopes from GPS profile and AirSWOT for the 499 overlapping 10 km reaches within the (a) west and (b) east regions versus the center of each reach in order of downstream distance. Successive reach segments are shifted downstream by 100 m.
3. Results
Figure 3: Differences between GPS profile and AirSWOT WSEs averaged over 1 km areas. Blue dashed lines mark the SWOT science requirement for WSE accuracies (±10 cm).
Figure 5: Histogram of slope errors for AirSWOT slopes relative to GPS profile slopes. Red dashed lines mark the SWOT science requirement for slope accuracies (±1.7 cm/km).
Figure 1: Tanana River study reach and AirSWOT elevation measurements from June 9, 2015. Grey water mask displays river extent not covered by AirSWOT measurements. GPS profile measurements (black line) and pressure transducer locations (blue triangles) are also shown.
East RegionWest Region
Flow Direction
2 2
2
2
We used a Trimble R9 survey-grade GPS system attached to the back of a 28-foot river boat to measure the in situ WSE profile (Figure 1).
Novel AirSWOT Measurements of River Height and Slope, Tanana River, AKElizabeth H. Altenau ([email protected]), Tamlin M. Pavelsky, Delwyn Moller, Christine Lion, Lincoln H. Pitcher, George H. Allen, Paul D. Bates, Stéphane Calmant, Michael Durand, Laurence C. Smith
University of North Carolina, Chapel Hill; Remote Sensing Solutions, Inc.; University of California, Los Angeles; University of Bristol; IRD/LEGOS Toulouse, France; The Ohio State University H21F-1483
Acknowledgements: This work was supported by NASA Terrestrial Hydrology Program Grant # NNX13AD05G managed by Jared Entin. We want to thank Curtis Chen, Gregory Sadowy, and Craig Stringham from the Jet Propulsion Laboratory and John Arvesen from Cirrus Digital Systems. Additionally, a special thanks goes to our boat driver Sam Demientieff.
References: • Biancamaria, S., D. P. Lettenmaier, and T. M. Pavelsky (2016), The SWOT Mission and Its Capabilities for Land Hydrology, Surv. Geophys., 37(2), 307–337, doi:10.1007/s10712-015-9346-y. • Rodriguez, E. (2016), Surface Water and Ocean Topography Mission (SWOT) Project - Science requirements Documents.
1 2 3 4 5 6
1 1 2 1 3 1 4 5 6 3