Icing challenges the safety and performance of wind turbines in low temperature environments. Detailed information on the amount and location of ice on rotor blades can facilitate decision making and mitigate hazardous effects. Therefore, a better understanding of the relation between ice accumulation and frequency response is required. Small scale experiments are conducted employing rotor blade specimens, experiencing natural icing of different configuration in terms of temperature and ice extent. The ice distribution is measured with a high-resolution laser scanner, and an algorithm based on alpha shapes is developed to determine the growth of the ice volume. Experimental modal analyses are performed periodically during the icing process, detecting consequent effects in the corresponding frequency spectra. The collected data is assessed regarding its suitability for an ice detection system and serves as a starting point for further developments.