Lightning hazard estimation by integrating surface electromagnetic and physical properties

We propose a method to estimate lightning hazard by integrating various physical surface properties and an electromagnetic parameter in order to present a lighting hazard map of northern Alberta, Canada. Physical surface properties include the land class, roughness, and temperature; whereas the electromagnetic parameter implies the estimated dielectric constant in this study. Geographic information system (GIS) data mining and spectral correlation methods are mainly carried out to estimate the potential lightning strike and consequent lightning hazard over the study area. The GIS data mining technique is implemented to find out the rule between the physical surface properties at each pixel and the lighting records. We compute the relative frequencies of the rules containing three different physical surface properties and sort them to identify which rule retains the highest possibility of lightning strikes. The potential lightning strike map is generated by normalizing the derived frequencies ranging from 0 to 1 and used with the non-hierarchical dielectric constant map in order to extract the pixels satisfying the condition of high dielectric constant and high frequency of a lightning strike by the wavenumner correlation filtering (WCF) method. The two maps filtered by the WCF are then combined by the local favorability index (LFI) to enhance the result. By correlating the potential lightning strike map with the non-hierarchical dielectric constant values in the spectral domain using the WCF and integrating them by the LFI, a lightning hazard of the study area is presented.