Using measured and derived aerosol properties from the Puerto Rico Dust Experiment (PRIDE), a four-stream broadband radiative transfer model is used to calculate the downward shortwave irradiance (DSWI) at the surface and the shortwave irradiance at the top of atmosphere (TOA). The results of the calculated DSWI are compared against pyranometer measurements from the Surface Measurements For Atmospheric Radiative Transfer (SMART) instrument suite at Roosevelt Road (18.20¿N, 65.60¿W). Using aerosol optical thickness retrievals from half-hourly geostationary satellite data (GOES 8 imager), the diurnal short wave aerosol forcing (SWARF) of dust aerosols both at the surface and TOA are calculated for the entire study area (14¿N ~ 26¿N, 61¿W ~ 73¿W). For selected days, the Clouds and the Earth Radiant Energy System (CERES) TOA shortwave irradiance values from Terra are compared with radiative transfer calculations. Wang et al. <2003> show that the satellite derived aerosol optical thickness is in excellent agreement with Aerosol Robotic Network (AERONET) values. Results of this study show that the calculated direct, diffuse and total DSWI are in excellent agreement with the corresponding SMART values with biases of 1.8%, -3.3% and 0.5% respectively, indicating that dust aerosols are well characterized in the radiative transfer model. This is well within the measured uncertainties (1.3%) and the model uncertainties (5%). The monthly mean value and standard deviation of aerosol optical thickness at 670 nm (AOT670) during PRIDE are 0.26 ¿ 0.13, and the corresponding monthly mean daytime SWARF values are -12.34 ¿ 9.62 W m-2 at TOA and -18.13 ¿ 15.81 W m-2 at the surface, respectively. Our results also show that if diurnal changes in aerosol optical thickness are not considered, it leads to uncertainties in SWARF of 4 W m-2 at the surface and 2 W m-2 at the TOA. The CERES TOA short wave irradiance underestimates calculated values by about 10 W m-2 mainly due problems in misclassification of aerosols and lack of aerosol angular dependence models (ADMs) in the current CERES algorithms. This study is among the first to demonstrate the potential of the GOES 8 imagers in retrieving aerosol optical thickness and estimating the daytime diurnal SWARF of dust, both at the TOA and surface, in low to moderate dust loading regions over the oceans.