I think calibrating the model again would be necessary. The reason is that rainfall gauges provide point data and satellite-rainfall products provide grid data with better spatial coverage.

@Alex the bias correction was applied on the satellite data in order to account for errors. Monthly accumulated rainfalls were computed for both satellite and rain gauge data and the satellite monthly rainfall is corrected by a factor to match the rain gauges and then disaggregated to daily rainfall

https://link.springer.com/article/10.1007/s13201-015-0330-y

hope this helps,

Please see

From your question I can understand that your model use the rain as input, and do not estimate it by itself. Therefore you need to relate to rain fall gauges data and satellite (or radar) which is always a tricky thing. You will need to calibrate the data for rain patterns differ from one place to another, and also at different meteorological conditions (wind direction, aerosol component etc). The calibration needs to meet your hydro-logical model time scale, weather is it a daily, monthly, yearly time scale model. Each will result in a very different calibration set. The longer the data sets used to calibrate the better calibration you get, and you can sub-divide it to different meteorological conditions. You might need to use statistical methods to overcome spatial or scale differences as well.

Best of luck

Thank you dear colleagues for your suggestions and responses. It sure did help me. 

You pose an interesting question, but I don't think we can provide useful answers without more specifics. For example:

These questions and others should be addressed before you make decisions on how to achieve your purpose. There is a lot of literature on the question of the validity of TRMM data and how to use it effectively in different situations. Some of which has already been cited above.

Thank you @Geoffrey those are important questions to ponder. Generally I have a moderate distribution of rain gauges (6 rain gauges in 800 km2), I am using MIKE NAM hydrological model and I am using an already calibrated model with in - situ precipitation data. My goal is to study the improvement/not of replacing in-situ precipitation data with satellite rainfall data. If its an improvement I want to use satellite data for some poorly gauged catchments and study how data assimilation can improve streamflow forecast with the use of satellite rainfall data. 

use two scenario

Check the quality of the TRMM data before you use them at all. We found very poor relations in our pixel-to-point validation, the TRMM products not even being able to capture the days with rainfall events, which make them quite useless in hydrological modelling. Have a look at attached publication.

Article Evaluation and comparison of satellite-based rainfall produc...