If you need more precise measurements of natural events on Earth's surface, get into space. Researchers studying glaciers and earthquakes have for some time followed this principle, exploiting the power of satellite interferometric imaging to map surface displacements down to the centimetre scale. Doug Alsdorf and his colleagues have taken the same approach in their investigations of the periodic floods that occur in the Amazon basin (D. Alsdorf et al. Geophys. Res. Lett. 34, doi:10.1029/2007GL029447; 2007).

Credit: P. BATES

The Amazon river has an intimate relationship with its vast floodplain, with an estimated 25% of its average annual discharge flowing and ebbing across it. But very little is known about the behaviour of these floods: not least, gauges of water level are placed only along the main channels, and then only sparsely. There are technical difficulties in taking interferometric measurements of water surfaces with satellite-borne synthetic-aperture radar. But flooded vegetation (pictured) does reflect an adequate signal.

Using data provided by instruments aboard the Japanese Earth Resources Satellite, Alsdorf et al. have been able to map the spatial and temporal complexity of floodplain inundation. Their study of floods from three different years takes in an area of the central Amazon basin that includes flows from the Purus river, as well as the Amazon itself.

Water levels in the floods, it turns out, do not take on the pattern that might be expected from a simple correspondence with the levels in the main channel of the river. Rather, there is a complicated interplay in which flow paths and water levels are influenced not only by the main channel and floodplain topography, but also by local and far-reaching hydraulic factors created by the flood itself.

These are proof-of-principle findings, with a practical edge. Modelling of floods is bedevilled by a lack of relevant measurements to test them. Satellite data can help redress that lack, with the ultimate aim of guiding engineering or other solutions to the inundation of areas inhabited by human populations. Furthermore, periodic flooding, and the associated delivery of sediments and nutrients, is a natural feature of wetland ecosystems not only in the Amazon but throughout the world. Some wetlands are under threat and, in some, restoration projects are in hand. Clarification of the relevant networks of water flow in different circumstances would offer another approach to ensuring the long-term success of such projects.