A scheme proposal for real time coherence tests of river water level derived from satellite altimetry

Nicolas Bercher (LEGOS/CNRS, France)


Stephane Calmant (LEGOS/IRD, France)

Event: 2014 Ocean Surface Topography Science Team Meeting

Session: Science Results from Satellite Altimetry: Inland waters (multi-mission and long-term monitoring)

Presentation type: Type Poster

The quality assessment of river water level time series derived from
satellite altimetry, by direct comparison to in situ gauging
measurements, has been addressed during the past two decades by
various research groups (Koblinsky 1993, Birkett 2002, etc.).
Ultimately, Cemagref has developed and implemented (2004-2008) a
standardised validation method prototype to automatise such
procedures. It gives synthetic quality indicators (Error RMS and
Sampling Loss Rate) at the scale of time series. The prototype has
been implemented on a wide extent of altimetry products (AVISO, ESA,
CASH, HydroWeb, River & Lake, PISTACH, CPP, etc.) from several
missions (T/P, ERS-2, Envisat, Jason-2). The quality assessment
exercise has been implemented for hundreds of virtual stations of the
Amazon basin, where in situ data are easily available. Results from
this prototype have shown progresses made by satellite altimetry for
the monitoring of river water level. They have been communicated on a
regular basis during altimetry meetings such as in Venice 2006 \& 2012
and Lisbon 2010.

However, in situ data usually have some important drawbacks, among
which the high latency of data availability imposed by measurement
system constraints (limnimetric scales read by human operator, digital
database filling from paper log books, verifications, validations,
etc.). This is really limiting for new missions such as SARAL/AltiKa
for which it is still not possible to perform validation on a complete
annual hydrologic cycle!

In such a context, were agencies and data users would like to get an
insight on the satellite measurements in real time or so, emerged the
need for tools to monitor measurement health, in real time, as new
data are acquired.

The proposed paper introduces a series of simple tests that helps to
monitor the coherence of any new incoming satellite measurement, in
near real time conditions and without the need for in situ data.
These tests are based on past measurements from the same location
(virtual stations coherence) as well as recent measurements acquired
by other missions on upstream or downstream tracks over the same
rivers (river profile coherence). Of course, past missions data can
also be reprocessed using the same tools.

Such tests highlights the need for high quality mean longitudinal
river profiles as well as the knowledge of mission/product specific

Future coherence tests might rely on basic river discharge
estimations, or on hydraulic and/or hydrologic models, at the scale of
a basin. Such approaches might also be integrated into time series
routines, used as filters prior to validation.

Nicolas Bercher