Better spatial resolution (smaller GSD) is more important in areas that are changing - like usually near coasts. Large, open areas of the ocean typically don't need high spatial resolution to study them: this is especially true in deep areas where the bottom has no impact on the signal, and where there is spatial uniformity on large spatial scales.  

Shallower areas where the bottom depth and bottom reflectance affect the returned signal benefit from higher spatial resolution. 

Improved spectral resolution is one way to improve the ability to discriminate the items that make up the remote sensing signal. However, one needs to understand that it is the combination of spectral resolution, spectral sampling interval and number of spectral bands (along with the spectral range sampled) that allow one to discriminate more components and/or more complex distributions. 

In general, there's likely no need for spectral resolution better than ~5 nm since most features of interest are larger than that. If a person wants to study certain Raman-scattering processes (in-filling of narrow solar absorption lines) then of course the researcher needs a much better spectral resolution. 

Some good papers on this are: Sensors 2007, 7, 3428-3441 Lee et al "Determination of Primary Spectral Bands for Remote Sensing of Aquatic Environments" and papers he references; another resource of interest is https://www.researchgate.net/publication/11368226_Effect_of_spectral_band_numbers_on_the_retrieval_of_water_column_and_bottom_properties_from_ocean_color_data

Finally, improved spectral resolution is relatively more important in coastal areas, too, since coastal areas are more complex than open-ocean areas. 

Article Effect of spectral band numbers on the retrieval of water co...

I totally agree with what Marcos says about spatial and spectral resolutions.  One comment I would add is that for shallow clear waters there will be 'some', but not much penetration of the water by NIR and SWIR spectral bands and most of the 'bottom' information will come from visible spectral bands.  In the deeper ocean waters of course there will be little to no penetration to the bottom and little spectral contrast in the more broad bands above the red (perhaps green) spectral bands.  Also an advantage of the moderate to lower spatial resolution (30m like Landsat TM to 250m like MODIS) images will be that individual pixels will not be influenced as much with glint caused by the waves.

Pat

I have raw input image with (15 m spatial resolution, and improving the  15 m spatial resolution to 10 m spatial resolution say for eg) . How this improved in spatial resolution have impact in near shore ocean for determining the bathymetry. 

Pat raises some good points. 

I will say that the glint/glitter/clutter issue is a bit more complicated. Overall, since larger pixels tend to have a mean flat surface, there tends to be little in the way of glint. However, even in AVIRIS 20 m data I can find glint, mostly due to the longer "swell" waves. And I also believe I have seen this in Hyperion 30 m data, too.

Jagalingam P:

It all depends on the spatial uniformity (or non uniformity) of the bottom; it probably also depends on the slope of the bottom; and it also depends on the means of determining the bathymetry. In general, better spatial resolution will allow better retrieval of the bathymetry. However, the details matter. What is the quality of the data? What spectral bands are present? What are they? What is your algorithm to determine the bathymetry? 

I was involved in a coral reef project in Hawaii and investigated the use of various remote sensing data sets to detect and help map the shallow coastal bottom.  Because of the features of interest high spatial resolution was needed, plus due to the depth some non-optical imaging was also used (airborne lidar and shipborne acoustic).  The aerial imaging included natural color with a spatial resolution of 10 to 15 cm and airborne lidar with 1m resolution (0.25m vertical resolution).  These are VERY clear coastal waters and the depth penetration varied from 5m to 45m depending on the spectral band (basically blue down to 45m, green to about 37m, and red to 5 to 7m); very little penetration with the NIR band.

The airborne lidar system had a blue/green wavelength laser which was able to penetrate down to about 50m and basically gave us a bathymetry data set very similar to a DEM on land.  Optical systems can give you some 1st order bathy information, but keep in mind that the 'scattering / bending' of the light by water is much worse than the scattering that occurs in the atmosphere.

I have attached a file with 3 slides showing an example of the results of the lidar image digitally over laid on the natural color aerial images. 

One more note and I'll stop rambling:

Keep in mind that the amount of 'glare or glint' from the sun can be minimized (either in shallow or deep waters) by pointing the imaging system slightly away from the sun, so with satellite imaging systems that have an off nadir pointing capability request images with a 5 to 15 degree off nadir view away from the sun.

Pat