Temporal variations in river water surface elevation and slope captured by AirSWOT

Abstract

25The Surface Water and Ocean Topography (SWOT) satellite mission aims to improve the 26 frequency and accuracy of global observations of river water surface elevations (WSEs) and 27 slopes. As part of the SWOT mission, an airborne analog, AirSWOT, provides spatially-28 distributed measurements of WSEs for river reaches tens to hundreds of kilometers in length. For 29 the first time, we demonstrate the ability of AirSWOT to consistently measure temporal 30 dynamics in river WSE and slope. We evaluate data from six AirSWOT flights conducted 31 between June 7-22, 2015 along a ~90 km reach of the Tanana River, AK. To validate AirSWOT 32 measurements, we compare AirSWOT WSEs and slopes against an in situ network of 12 33 pressure transducers (PTs). Assuming error-free in situ data, AirSWOT measurements of river 34 WSEs have an overall root mean square difference (RMSD) of 11.8 cm when averaged over 1 35 km 2 areas whilst measurements of river surface slope have an RMSD of 1.6 cm/km for reach 36 lengths >5 km. AirSWOT is also capable of recording accurate river WSE changes between 37 flight dates, with an RMSD of 9.8 cm. Regrettably, observed in situ slope changes that transpired 38 between the six flights are well below AirSWOT's accuracy, limiting the evaluation of 39 AirSWOT's ability to capture temporal changes in slope. In addition to validating the direct 40 AirSWOT measurements, we compare discharge values calculated via Manning's equation using 41AirSWOT WSEs and slopes to discharge values calculated using PT WSEs and slopes. We 42 define or calibrate the remaining discharge parameters using a combination of in situ and 43 remotely sensed observations, and we hold these remaining parameters constant between the two 44 types of calculations to evaluate the impact of using AirSWOT versus the PT observations of 45 WSE and slope. Results indicate that AirSWOT-derived discharge estimates are similar to the 46 PT-derived discharge estimates, with an RMSD of 13.8%. Additionally, 42% of the AirSWOT-47 based discharge estimates fall within the PT discharge estimates' uncertainty bounds. We 48 conclude that AirSWOT can measure multitemporal variations in river WSE and spatial 49 variations in slope with both high accuracy and spatial sampling, providing a compelling 50 alternative to in situ measurements of regional-scale, spatiotemporal fluvial dynamics. 51

A High-Resolution Airborne Color-Infrared Camera Water Mask for the NASA ABoVE Campaign

“…A key SWOT mission objective is to obtain water surface elevation (WSE) to ±10 cm vertical accuracy for 1 km 2 open water regions [66]. Previous work used CIR-derived water masks and spatially-averaged AirSWOT Ka-band interferometric radar data to estimate AirSWOT WSE accuracies of~9-10 cm in rivers and~21 cm in lakes [53,[67][68][69]. The conservative AirSWOT color-infrared (CIR) water mask provided here should enable unambiguous extraction of open water pixels used for spatial averaging of AirSWOT interferometric radar data, and, therefore, improve estimates of WSE accuracy (Figure 10).…”

Mapping Water Surface Elevation and Slope in the Mississippi River Delta Using the AirSWOT Ka-Band Interferometric Synthetic Aperture Radar

“…Similarly, Tuozzolo et al [30] used AirSWOT data to estimate WSE, WSS, and river discharge over the Willamette River in Oregon, USA, reporting WSE and WSS RMSE of 11.6 cm and 3.2 cm/km, respectively, and discharge error of 10 to 31% depending on discharge algorithm and accurate prior estimate of mean annual discharge. While these uncertainties are slightly higher than that of Altenau et al [27,28] or Pitcher et al [29], Tuozzolo et al noted the Willamette is a thinner river, requiring smaller averaging windows (i.e., less pixels). Appropriate methods for averaging and filtering of AirSWOT data enable accurate estimation of WSE and WSS in spite of the noise in AirSWOT measurements [27][28][29][30][31].…”

“…While these uncertainties are slightly higher than that of Altenau et al [27,28] or Pitcher et al [29], Tuozzolo et al noted the Willamette is a thinner river, requiring smaller averaging windows (i.e., less pixels). Appropriate methods for averaging and filtering of AirSWOT data enable accurate estimation of WSE and WSS in spite of the noise in AirSWOT measurements [27][28][29][30][31].…”

“…Early studies had demonstrated the feasibility and potential of Ka-band InSAR for mapping water surfaces, using both ground-based [24] and airborne radar experiments [25,26]. Recent studies have demonstrated the capabilities of AirSWOT to measure rivers and lakes [27][28][29][30][31]. Altenau et al [27] used AirSWOT data to measure WSE, WSS, and estimate discharge [28] along the Tanana River, Alaska, USA.…”

“…Recent studies have demonstrated the capabilities of AirSWOT to measure rivers and lakes [27][28][29][30][31]. Altenau et al [27] used AirSWOT data to measure WSE, WSS, and estimate discharge [28] along the Tanana River, Alaska, USA. They reported WSE root mean square error (RMSE) of 9.0 cm over 1 km 2 areas, WSS RMSE of 1.0 cm/km along 10 km river reaches, and discharge RMSE around 10-20%.…”