Location Accuracy Improvements in the Japanese Lightning Detection Network by Eliminating the Data Obtained from Distant Sensors from Lightning Stroke Positions

Abstract:

Authors have observed the current waveform of lightning flashes by the Rogowski coils installed at the wind turbines located in southern Kyushu, Japan. We succeeded in observing three sets of lightning current waveforms of negative downward flashes hit to the wind turbine and estimated the location error by comparing the actual position with the positions located by the Japanese Lightning Detection Network (JLDN) in two summer seasons between 2013 and 2014. The JLDN is a Lightning Locating System consisting of 30 LF sensors installed for observing lightning in the whole area of Japan. According to our estimation, the mean location error of the first strokes was 557m and that of subsequent ones was 316 m in the JLDN, respectively. This means the location accuracy of the JLDN was slightly less than that of the NALDN and the EUCLID. We found out location errors were less than 400 m when only information from sensors within 700 km from lightning strokes was used for locating stroke positions. Therefore, we recalculated locations of those lightning strokes without information from sensors at more than 700 km from them. In the results of the recalculation, the mean location error of the first strokes was improved by 219 m and that of subsequent ones was improved by 48 m, respectively. This indicates eliminating large deviation data from distant sensors for lightning location reduced the standard deviation of the time differences between actually measured time at sensors and calculated time and thus improved location error of the JLDN. We found out that there were strong positive relations between the location errors and the standard deviation of the time differences. The standard deviations of the time difference were less than 1.5 micro-seconds when location errors were within 400 m. Therefore, locating lightning strokes without information from distant sensors is a useful method for reducing the standard deviation and improving location error.