DREAMing in River Basins

The contribution of  satellite radar altimetry to river monitoring is well-established, with data forming  valuable inputs to river models.
Surface soil moisture can also be determined from altimetry using DRy EArth Models (DREAMs) which model the response of a completely dry surface to Ku band nadir illumination. New DREAMs over Africa now cover more than 70% of the continent, encompassing more than 30 river basins including the Congo, Niger, Okavango, Zambezi and Volta.  

It was decided to fly multi-mission altimetry over these river DREAMs to assess the potential of this technique to contribute to studies in river basins. As a detailed DREAM exists for the Amazon basin, this was also included. Envisat, ERS-1/2, Jason-1/2, CryoSat-2 and Sentinel-3A altimeter data were utilised in this study, together with a database of over 86000 graded altimeter River and Lake height time series. Soil moisture estimates were generated and validated.

Summative conclusions: the highest data retrieval rate over river DREAMs is found over ‘river’ and ‘wetland’ pixels, with lower percentages over ‘soil’ pixels where soil moisture estimates can be generated. This is an expected outcome, as targeting ‘soil’ pixels will select for rougher topography. 
Within the constraints of satellite orbit and repeat period, data can be successfully gathered over the majority of these overflown DREAM surfaces. It is also clear that very detailed DREAM models, at least 10 arc seconds resolution, are required to capture the intricate structure in river basins. It is noted that many tributaries are below the current 10 arc seconds spatial resolution of the DREAMs, and are classified with their surrounding terrain as wetland pixels.
ERS-2 and Envisat performed best; Sentinel-3A OLTC mask is found to preclude monitoring of almost all ‘soil’ pixels, except those adjacent to the largest rivers.
The ability of nadir-pointing altimeters to penetrate vegetation canopy gives a unique perspective in rainforest areas. Amazon soil moisture time series in the lower Amazon are seen to correlate to river height variations: in the upper Amazon basin the annual rainfall signature is dominant.
Over much of the river DREAMs, along-track time series of soil moisture can be generated at the spatial resolution of the underlying DREAM, currently 10 arc seconds.  The major constraint, as with altimeter height measurements, is the spatio-temporal sampling, so use is envisaged in combination with other remote sensed and in-situ data.  However, DREAMing provides a valuable independent dataset which can be used to validate soil moisture estimates from other techniques.