Assuming your experimental setup determines the height of the river at some point, one can compute the discharge directly from other known hydraulic parameters (see http://onlinelibrary.wiley.com/doi/10.1002/2015WR018434/full, which is attached).

I am attaching a pamphlet on discharge measurements that may help.  I would suggest if nothing else installing one or more transducers (they are relatively inexpensive, can detect water level changes, and record data at almost any frequency).  They can however foul on occasion if too much sediment.  Developing a stage (water level) discharge is typically a time consuming endeavor, and for large rivers, with some hazards requiring quality equipment. The acoustical doppler meters are what you probably need to help measure flow velocities during cross section measurements at different river stages, however, these are fairly expensive.  They too may have some issue with high sediment loads if that is a problem.  If bridges over river are available, the old methods with current meter still work to help determine the velocity profiles at points across the cross section of the river.  Cable cars such as those on document cover are sometimes used if there are no reasonable crossings if the intent were to use the old style current meters.  

I am not sure how good the present optical remote tools might be or how coarse of an estimate you might be willing to accept.  

Basically a detailed cross section of the river that is stable in its channel properties (not aggrading, not degrading, not shifting location) is needed, and a total station survey of channel levels and floodplain levels can give a high quality cross section of the channel and floodplain to uplands that will not flood.  This is needed so as river stage changes, you can determine cross section change, and with mean velocity (a composite of many measures), you can estimate discharge changes.  

There are other ways to measure velocity, with substantial variance in accuracy.  One old school method to estimate velocity was to throw in several oranges and time how long they take to float a distance.  

If other systems are not reasonable.  Discharge can also be estimated by injecting at a known concentration of a fluorescent dye at a known constant rate, and measure the concentration at a well mixed station downstream until the concentration stabilizes downstream to determine the dilution factor, and multiply this by the injection rate.  The objective is typically to develop the stage discharge relationship of a stable section, so water level changes detected can be used to estimate discharge.

Please have a look at these useful PDF attachments.

Dear Shahzad, look at the PDF in attached

Dear All

Thank you so much for providing useful information and guidance!

Dear Dr. Sultan, 

If you have access to a series of images of your location of interest, from which a time series of river widths can be reasonably estimated, you could apply a method called at-many-stations to estimate discharge. This method, which is one of the methods evaluated in the article that Dr. Hunt attached, uses scaling laws between river width and discharge to estimate discharge.  The full description of the method is presented in the attached article entitled “Toward global mapping of river discharge using satellite images and at-many-stations hydraulic geometry” authored by Prof. Gleason.

The at-many-hydraulic-stations method was developed to work for large areas taking advantage of satellite imagery.  If instead what you intend to use are video images of a particular location to estimate discharge, a technique that could be applied is based on large scale particle image velocimetry. This is not my area of expertise, but I remember a prototype of a LSPIV system that was built in the laboratory where I worked during my PhD work.  It is simple to build and operate and compared well with a collocated gauge during tests.  If that’s the avenue that you would like to pursue, you could find details on the Kim et al (2008): Stream discharge using mobile large-scale particle image velocimetry: a proof of concept (http://onlinelibrary.wiley.com/doi/10.1029/2006WR005441/pdf). If you’d like a copy, please message me, as I’m unsure if that article’s copyright allows open sharing.

Best regards,

Renato Frasson

Article Toward global mapping of river discharge using satellite ima...

i agree with Hansen

Thank you Renato Frasson

I believe the concept you shared is very useful and possibly we could try to apply on River Indus.

Thank you all