Precipitation in the Caribbean basin can be characterized into three rainfall seasons; the Dry Season (DS), December-April, the Early Rainfall Season (ERS), May-July, and the Late Rainfall Season (LRS), August-November with a brief dry period in July exhibiting a climatological bimodal behavior. Sea surface temperatures and Vertical Wind Shear (VWS) have been reported as the main drivers for the precipitation in the DS and the LRS, respectively, while little is known about the ERS and the summer drought. It has been hypothesized that increases in giant aerosols concentration due to the Saharan Dust across the Caribbean in the summer months may result in precipitation suppression. A multivariable analysis was carried to determine which climatological variables may correlate with the Caribbean summer drought that included Intertropical Convergence Zone (ITCZ), the North Atlantic Oscillation (NAO) index, the Vertical Wind Shear (VWS) and the aerosol particles coming from northern Africa. The ITCZ is calculated using the SST data from Reynolds-Smith NOAA-NCEP data set, NAO is obtained from Climate Prediction Center-National Weather Service, the VWS is calculated from the NCEP-reanalysis data and the aerosol particles (AP) is acquired from Aerosol Robotic Network (AERONET). The analysis shows that the ITCZ and the SST are weakly correlated with the Caribbean bimodal precipitation; however, the VWS and aerosol particles revealed an important contribution to rainfall during the summer months. The multiple regression analysis also reveals an independent strong correlation of the VWS. Spectral analysis revealed that the NAO does not have periodicity leaving the VWS and AP as potential controlling variables. Numerical experiments are performed to uncouple the VWS and AP effects in the lower summer rainfall time. The numerical approach uses the regional atmospheric modeling system (RAMS) with a new cloud microphysics module which considers small and giant AP. A control run is carried out for July 2003 which has a VWS closer to the climatology and a second run for July, 2002, where a VWS anomaly is present. These numerical experiments supported the statistical result that the VWS effect uncoupled to the AP influences the rainfall production in July. Two additional numerical experiments are performed in which observed AP are ingested into RAMS. The first of these simulations uses observed Cloud Condensation Nuclei (CCN) and Giant CCN measured in July 2003, and in a fourth run the concentration of AP are increased maintaining the same atmospheric conditions as in the previous run. Results indicate that higher AP concentrations decreases further rainfall in July across the Caribbean. Possible variability of this bimodal trend with climate changes is also explored in the paper.