We report (quasi) periodic oscillations observed in soil radon emanation data especially during summer period. Soil radon has been continuously monitored in the Marmara region of Turkey over the past nine years to reveal possible relationships between soil radon and seismic activities. This long term monitoring has clearly demonstrated that soil radon concentrations are affected by various parameters such as seasonal and daily changes in atmospheric parameters (temperature, pressure, precipitation). Sometimes, soil temperature variations as well as barometric pressure and precipitation may dominantly influence the soil radon concentration; leading to seemingly complex radon time series. One may need to remove such components for better analysis of the possible relationships between soil radon emanation and seismic activities. Here, we suggest using an algorithm to detect and remove daily-oscillations from the raw data. The detection and separation (extraction) algorithm is based on Empirical Mode Decomposition method which is a signal-adaptive method that automatically decompose the signal into its characteristic modes. The algorithm is applied to the data collected from three different soil radon monitoring stations (Balikesir, Gonen, and Armutlu) in Marmara region of Turkey and the results are provided. (C) 2010 Elsevier Ltd. All rights reserved.