OPERATIONAL HELIOSPHERIC SPACE WEATHER MODELS REQUIRE SIMULATIONS OF CORONAL MASS EJECTIONS (CMES) MUCH FASTER THAN REAL TIME FOR ALL OBSERVED CME EVENTS. CURRENTLY THIS CANNOT BE ROUTINELY ACHIEVED BY SIMULATIONS THAT INVOLVE THE INITIATION OF CMES NEAR THE SUN. THE WSA-ENLIL-CONE MODEL THAT USES GEOMETRIC AND KINEMATIC PARAMETERS OF CMES OBSERVED IN CORONAGRAPHS AND LAUNCHES SIMPLE HYDRODYNAMIC CME-LIKE STRUCTURES INTO THE BACKGROUND SOLAR WIND. PROPAGATION AND INTERACTION IN THE HELIOSPHERE ARE SUBSEQUENTLY SOLVED BY A 3-D MAGNETOHYDRODYNAMIC (MHD) CODE. THE WSA-ENLIL-CONE MODEL HAS BEEN ROUTINELY USED AT NASA AND NOAA. VALIDATION STUDIES HAVE FOUND THAT CME ARRIVALS CAN BE PREDICTED WITH AN UNCERTAINTY OF ABOUT 6-9 HOURS. THIS IS PROBABLY THE BEST PERFORMANCE THAT CAN BE ACHIEVED FOR FIXED VALUES OF THE CONE MODEL FREE PARAMETERS (LIKE THE EJECTA RADIAL EXTENT AND THE DENSITY ENHANCEMENT) THAT CAN BE DETERMINED FROM CALIBRATION STUDIES. TO FURTHER IMPROVE THE PREDICTION ACCURACY OF CME ARRIVALS WE PROPOSE TO ENHANCE THE MODEL CAPACITY THROUGH THE FLEXIBLE USAGE OF THE FREE PARAMETERS THAT CAN BE CONSTRAINED BY EXISTING OBSERVATIONS. WE WILL USE THE WHITE-LIGHT HELIOSPHERIC OBSERVATIONS OF THE CME ACCELERATION/DECELERATION AND AN EMPIRICAL MODEL OF THE CME PROPAGATION WITH VARIABLE DRAG COEFFICIENT TO ESTIMATE THE INITIAL DENSITY OF THE PARTICULAR EJECTA. THIS PROCEDURE WILL BE FURTHER REFINED BY COMPARING THE OBSERVED AND SIMULATED TIME-HEIGHT PROFILES OF THE CME LEADING EDGE. THE PROPOSED PROJECT COMBINES EXISTING OBSERVATIONS AND MODELS TO INCREASE THE PREDICTION ACCURACY OF CME ARRIVALS IN THE WSA-ENLIL-CONE MODEL. THE RESULTING PRODUCT WILL BE DELIVERED TO NASA AND NOAA.