Previous studies have suggested a possible association between meteorological factors and the onset of subarachnoid hemorrhage (SAH). We aimed to investigate the relationship between the onset of SAH and meteorological factors based on an hourly time-series analysis. We collected hourly data on transportation of patients with SAH using the ambulance records of the Tokyo Fire Department from January 1 to December 31, 2005. We also collected hourly meteorological data for Tokyo from the Japan Meteorological Agency during the same period. We performed a time-series analysis using the autoregressive integrated moving average (ARIMA) model to control for autocorrelations in the time-series data. There were 1729 patients with SAH (mean age 63.3 years; 60.2% women). We identified two circadian patterns in the onset of SAH: a daily peak at 10 am (p<0.001) and a seasonal peak in February (p<0.001). Based on the ARIMA time-series analysis, significant risk factors associated with the onset of SAH included: low temperature on the previous day (lag time 17h; p=0.005) and on the onset day (lag time 0h; p<0.001); high barometric pressure on the onset day (lag time 0h; p=0.001). Humidity was not associated with the onset of SAH. Among meteorological factors, low temperature and high barometric pressure may be risk factors for the onset of SAH.