Abstract: In order to determine optimal calculation method for identification of abnormal temperature areas, firstly, three commonly used regional data processing methods (traditional statistical method, trend surface analysis method and fractal method) were applied to process radon measurement data of actual mine surface. Then, differences of three methods were compared and analyzed from ability of strengthening weak anomalies and determination of outlier threshold. Finally, representative boreholes were selected to verify anomaly areas determined by the three methods. The results show that the lower limits of radon anomaly of traditional statistical method and fractal method are 5 982 Bq/m³ and 7 094 Bq/m³ respectively, and trend surface analysis method is used to determine anomaly area by positive residual component. In delineating abnormal areas according to its consistency with ability of strengthening weak anomalies, the three methods rank as trend surface analysis method > fractal method > traditional statistical method. By comparing results of drilling verification, it is determined that results delineated by fractal method can be used as basis for later fire-fighting activities.

Key words: isotopic radon measurement, abnormal areas, fire areas zoning, underground fire source detection, fractal method