By using an estimate of residual noise (i.e., LA90+2dB), the level and variability of residual noise were assessed for sites of mid-/long-term or year-round observation. The sites were classified into rural, residential, urban, or shopping area. The level and variability of residual noise were determined by year, month, and day of week. Analysis of variance clarified the relationship between various elements. Short-term (one-week) observation was implemented for many additional sites. It was found that 1) in rural areas, the level of residual noise is large variation on the time of year; in urban areas, it is small variation on the time of year. In general, variance of residual noise is large at sites where the noise level is low. Therefore, it is difficult to obtain reliable measurements at quiet sites; 2) measurement of residual sound at rural sites is affected by insect and frog sounds from early summer through autumn, although reliable measurements are ensured from November through April at these sites; and 3) because measurement at urban and shopping sites is affected by commercial and industrial activities, weekday measurements should be evaluated separately from holiday measurements, for accuracy.