Hi - my first suggestion would be for you to define clearly your "problem". After all, you can use SAR images to answer a number of flood related questions. I assume you are interested in maximum flooding extent. In principle, if there is a temporal offset between your images(s) and the max flood extent you have a problem. If you have additional information, for example whether the peak occurred before or after the image acquisition date, you can refine your answer. If the image acquisition took place after the max peak, depending on how much time has passed, the (likely still) high soils moisture, or the relative smoothening of the area that was flooded can still help you to delineate the peak water extent like. Often water also leaves a bit of a debris line that should be detectable in the radar imagery. If, however, the image was acquired before the peak was reached only a modelling approach can help you, where continued water input into the system, and detailed elevation, needs to be used.

Hope this is useful, Norman

Thank you for your suggestion. Definitely, it is very helpful. 

The advice provided is appropriate.  There, are various other tools a hydrologist may apply, and field determinations, debris or other high water indicators can often be read with some skill.  If in a developed or urban area, high water marks should be evident, and flooding extent may be estimated from the drop in the SAR elevation of water from georeferencing high water marks as indicator points, and then use the GIS DEM to extend the elevation indicators back to high water marks.  You may even find flood debris in trees or brush. 

In addition, streams have generally as defined fall rate from a peak that if known, that recession curve might be able to be extended backward in time to an approximate peak.  Sometimes there may be equations to help estimate peak flows at ungauged sites, from then gauged stream record.  Several methods might be tried and then look at the various evidence to come to an approximation.