We estimated monthly atmospheric turbidity from hourly observation data of global solar radiation under cloudless skies at 67 stations all over Japan. Using an empirical formula for calculating global solar radiation, we determined atmospheric turbidity as the best-fit parameter. The estimated turbidity is consistent with the observation by the Japan Meteorological Agency. Sub-nationwide increases in atmospheric turbidity, caused by forest fires or Asian dust events, are detected in time series. Some temporal changes in atmospheric turbidity are also found in a small area. Such local phenomena are hardly detectable by sparse distribution of the stations for direct solar radiation observation. The accuracy of the global solar radiation estimation can be improved by considering the seasonal variation of the turbidity. Since Ångström’s turbidity coefficient is directly linked with the volumetric aerosol content, the present method is useful for monitoring the spatial distribution of aerosols. In addition, the present method using the densely distributed observation network of global solar radiation has an advantage in compensating for the sparseness of the current atmospheric turbidity observation in Japan.