Yes of course.. Calibration and validation are necessary for modelling and to be done to get the accurate results.. Outcome from your modelling work will depends on the measured data..

Definitely, you need to calibrate and validate the model to reliable results. The calibrated data shod be different than the validation data.

Not necessarily measured. In another word, some parameters may be estimated using resression and statsitical procedures. Yet the accuracy will indeed improve with more measured parameters.

Of course, parameter estimation or calibration is a key step for modeling. But it does not mean that all the parameters should be calibrated using computer procedures. I think only these important parameters should be calibrated, while those non-sensitive parameters can be fixed at normal values.

You need to calibrate only necessary parameters which are relevant to the main water quality parameter you want to model+ all hydraulic parameters

A senstivity analisys would help you to find out, which parameters are important, so maybe you can reduce the number of parametes you have to use

We should clearly know what calibration and validation are before answering this questions. Model calibration is concerned with the process of estimating model coefficients and constants. Model validation is concerned with building right model, which tests the accuracy of the model’s representation of the real system.

Hence it depends on the methods you use for calibration. For example, if you get a parameter (a) from a ratio of two observed data (x,y), i.e. a=y/x, y=a*x, or from a physical law, then it is unnecessary to validate it.

You first need to define which variables you want as output to answer your question. You then perform a sensitivity analysis of these variables with respect to the model parameters. Most likely a number of parameters will turn out not to be influential (at all!) and it is then no use to try to calibrate them. It is better to keep them then on default values since these values represent our domain knowledge. Most of the time the hydraulic parameters affect the water quality variables significantly and will therefore be part of the set of parameters to be calibrated. Depending on the quality of the measurement set you have, you will be able to estimate more or less of the set of influential parameters you determined by the sensitivity analysis. Good luck!

HECRAS is almost a Semi-2D model since the cross section is simulated by an average depth based on the submitted data for the x-sec.  It will be useful if you want to know the trend or the behavior of the WQ parameter in the downstream.  It  will not give you any information in the cross direction.  Thus, it will be very difficult to simulate plume pollution for example since it is an explicit  2-d problem.  

Calibration will be very important to identify the behavior of  te reach you are working on.  But you must be sure that the problem you are solving can be simulate via 1-d modelling approach.