ASSIMILATION OF PRECIPITATION HAS BEEN GENERALLY UNSUCCESSFUL BECAUSE: A) PRECIPITATION IS NOT GAUSSIAN AND THIS VIOLATES THE BASIC ASSUMPTION OF CURRENT DATA ASSIMILATION METHODS AND B) FORCING THE MODEL TO RAIN WHERE OBSERVED BY MOISTENING OR DRYING THE ATMOSPHERE AS DONE IN DATA ASSIMILATION IN THE PAST DOES NOT CHANGE THE MAIN DYNAMICAL VARIABLE (POTENTIAL VORTICITY) AND THUS IS IMMEDIATELY FORGOTTEN AS SOON AS THE FORCING OF THE "ASSIMILATION OF MOISTURE" CEASES. HOU ET AL. (2004) REPORTED THE MOST SUCCESSFUL CASES OF IMPROVING FORECASTS BY ASSIMILATING PRECIPITATION IN THE PRESENCE OF HURRICANES. IN THESE CASES THE FORCING OF THE LARGE INCREASED PRECIPITATION WAS ABLE TO MODIFY THE POTENTIAL VORTICITY OF THE MODELED HURRICANES AND THUS SIGNIFICANTLY IMPROVE THEIR FORECASTS. LIEN ET AL. (2013 2016A 2016B) AND LATER KOTSUKI ET AL. (2017) DEVELOPED A METHODOLOGY THAT SUCCESSFULLY ADDRESSED THESE PROBLEMS: THEY PERFORMED A TRANSFORMATION OF OBSERVED AND MODELED PRECIPITATION INTO GAUSSIAN VARIABLES WITH GAUSSIAN ERRORS (ANAMORPHOSIS). THEY ALSO USED ENSEMBLE-BASED DATA ASSIMILATION APPROACHES THAT ARE MORE NATURALLY ABLE TO ASSIMILATE PRECIPITATION AND DIRECTLY INFLUENCE THE MODEL DYNAMICAL VARIABLES I.E. POTENTIAL VORTICITY SINCE THE ENSEMBLE MEMBERS THAT PRECIPITATE CLOSELY AS OBSERVED RECEIVE HIGHER WEIGHTS IN THE ENSEMBLE ANALYSIS AND THEREFORE "DONATE" THEIR MORE CORRECT DYNAMICS TO THE ANALYSIS MEAN THUS CREATING AN ANALYSIS MEAN THAT IS CLOSER TO HAVING THE RIGHT DYNAMICS. AS A RESULT BOTH LIEN ET AL (2016) AND KOTSUKI ET AL (2017) SHOWED THAT IT WAS POSSIBLE TO ASSIMILATE REMOTELY SENSED PRECIPITATION (NASA TMPA AND JAXA GSMAP) AND IMPROVE THE FORECAST SKILL FOR OVER 5 DAYS. THROUGH THE COLLABORATION WITH RIKEN JAPAN WE APPLIED SIMILAR METHODOLOGY TO THE JAPAN REGIONAL MODEL SCALE AND FOCUS ON IMPROVING THE HURRICANE FORECASTS. OUR PRELIMINARY RESULTS SHOW THAT ASSIMILATION OF THE JAXA GSMAP PRECIPITATION A PROXY OF NASA IMERG IMPROVES BOTH THE TRACK AND INTENSITY FORECAST OF HURRICANES. MEANWHILE WE FOUND SEVERAL DEFICIENCIES OF THE CURRENT PRECIPITATION FRAMEWORK ESPECIALLY THE STATISTICAL METHODOLOGY USED BY LIEN ET AL. (2016B) TO PERFORM THE GAUSSIAN TRANSFORMATION ALTHOUGH OPTIMAL FOR GLOBAL USE IS BASED ON CLIMATOLOGICAL ASSUMPTIONS THAT ARE NOT APPLICABLE TO EXTREME EVENTS LIKE HURRICANES. HERE WE PROPOSE TO USE THE METHODOLOGY DEVELOPED BY LIEN ET AL (2016A B) AND MODIFY IT TO FOCUS ON IMPROVING THE FORECASTS OF HURRICANES. SPECIALLY WE PROPOSE TO: A) DEVELOP NEW GAUSSIAN TRANSFORMATION METHODS SUITABLE FOR TRANSFORMING HURRICANE RAINFALLS AND ASSIMILATING NASA IMERG PRECIPITATION OVER OCEAN TO IMPROVE THE HURRICANE PREDICTION. B) TRY TO USE THE STATE-OF-ART TECHNIQUE ENSEMBLE FORECAST SENSITIVITY TO OBSERVATIONS (EFSO) AS THE GUIDELINE TO ASSIMILATE IMERG RETRIEVALS OVER LAND WHERE THOSE RETRIEVALS SUFFER LARGE UNCERTAINTIES AND SHOW LIMITED IMPACTS ON IMPROVING FORECASTS SKILLS C) PHYSICALLY AND SYSTEMATICALLY DETERMINES THE PARAMETERS RELATED TO PRECIPITATION ASSIMILATION BASED ON THE EXPERIENCES SHARED BY THE PRECIPITATION RETRIEVAL COMMUNITY TO MAXIMIZE THE BENEFITS OF PRECIPITATION ASSIMILATION AND D) EXPLORE IF PERFORMING HIGH-FREQUENCY ASSIMILATION CYCLE WITH THE UNIQUE HIGH-TEMPORAL-RESOLUTION (30MINUTES) IMERG WILL FURTHER ENHANCE THE INTENSITY FORECASTS OF HURRICANES.