GEOPHYSICAL RESEARCH LETTERS, VOL. 32, L17404, doi:10.1029/2005GL023836, 2005
Water slope and discharge in the Amazon River estimated using the shuttle radar topography mission digital elevation model
Gina LeFavour1 and Doug Alsdorf 2
Received 16 June 2005; revised 21 July 2005; accepted 8 August 2005; published 8 September 2005.
[1] We find that the standard deviation, hence error, of the water surface elevation data from the Shuttle Radar Topography Mission (SRTM) is 5.51 m for basin-wide, regional and local Amazon mainstem reaches. This error implies a minimum reach length of 733km in order to calculate a reliable water-surface slope. Resulting slopes are 1.92 ± 0.19 cm/km for Manacapuru, 2.86 ± 0.24 cm/km for Itapeua and 3.20 ± 0.34 cm/km for Tupe. Manning's equation is applied with these slopes and with channel width measurements from the Global Rain Forest Mapping project synthetic aperture radar mosaics (GRFM SAR), channel depths averaged from nautical charts, and reasonable estimates of Manning's n. Resulting discharge values are 84,800 m3/s at Manacapuru, 79,800 m3/s at Itapeua, and 62,900 m3/s at Tupe averaged over the SRTM mission period. These values are within 6.2% at Manacapuru, 7.6% at Itapeua, and 0.3% at Tupe of the insitu gage-based estimates for the same or similar time period. Citation: LeFavour, G., and D. Alsdorf (2005), Water
slope and discharge in the Amazon River estimated using the shuttle radar topography mission digital elevation model, Geophys. Res. Lett., 32, L17404, doi:10.1029/2005GL023836.
[3] Remote sensing has the potential to offer the spatial and temporal coverage necessary to measure water fluxes globally [e.g., Alsdorf et al., 2003; Alsdorf and Lettenmaier, 2003]. Most water flux estimation methods use regressionbased relationships between remotely measured inundated area and in-situ measured discharge to predict fluxes [e.g., Smith, 1997]. However, this approach does not work well in environments where small changes in water heights yield little change in surface area yet significant changes in flow. For example, an elevation change of 5 cm over 24 hours in just 1/25 of the Amazon's flooded 750,000 km2 can be equivalent to the average discharge of the Mississippi River, yet the change in flooded area is practically undetectable at any spatial resolution. Instead, Alsdorf and Lettenmaier [2003] have called for methods that measure surface water hydraulics, particularly elevations, to estimate Q and S. In this paper, we provide the first documented use of the SRTM DEM to estimate stream channel water elevations, slopes, and discharges (C-band, 3 arc second data are used).

下一页