Error Analyses and Network Optimization for Time-of-arrival Lightning Locating System

Time of arrival (TOA) measurements in lightning studies is pioneered by Proctor D E in South Africa. With the high accuracy and large detection range, TOA lightning locating system are widely used all over the world. A variety of studies are made on the retrieval of lightning locations and analysis of location errors. Results show that network geometry, number of antennas, differencing scheme, timing error are important factors affecting retrieval accuracy. Location errors of TOA lightning locating system are analyzed, and effects of station amount, network geometry and baseline length on location accuracy are investigated too. A linear algebraic algorithm is introduced for retrieving the location and time of occurrence of lightning ground strikes from a TOA network. Based on the algorithm, Monte Carlo computer simulation method is used to estimate the locating error, assuming that the system has an overall timing root mean square error of 1 μs , neglecting all other possible errors (e.g., anomalous noise). The detailed spatial distributions of retrieval errors are provided. Results of this research provide scientific basis for design and optimize a lightning network in practical work. The results indicate, in a lightning locating system, location errors do not simply decrease with the amount of stations. They are closely related with network geometry. As for geometry of rectangle, diamond, star and triangle shape, rectangle network with a central station have better location accuracy. In a lightning locating system, sources outside the network have lager location errors than that inside the network. Range errors outside the network increase with distance away from center of the network. With a certain network geometry, location errors decrease with station amount. Regional lightning locating system requires a minimum of four stations. A good network requires three or more widely spaced stations transverse to waves approaching from any direction. A square network of four stations has blind regions near the symmetry axes. This problem can be solved by adding a central station in the square. Within the effective detection range of the network, coverage area and location accuracy both increase with the length of the baseline. On determination of baseline length, factors such as antenna's detection efficiency, timing errors, terrain are all important variables that should be considered.