Vaisala Global Lightning Dataset GLD360: Performance, Improvements, and Plans

For over two years, Vaisala has been operating a long-range lightning geo-location network designed to provide high precision ground flash data, the Global Lightning Dataset GLD360, at global distances. Peak current and polarity estimates are given for each event, a unique feature among long-range lightning detection networks. This paper presents a spatio-temporal statistical comparison between individual GLD360 events and cloud-to-ground strokes detected by the well-characterized U.S. National Lightning Detection Network® (NLDN®). The new processing algorithm brings the overall median location accuracy to between two and five km, depending on the time of day. Peak current estimation error is 10 to 30%, also depending on time of day, and percent correct polarity estimation is nearly 90%. The overall cloud-to-ground stroke-level detection efficiency (DE) is near 60%, with higher DE rates for strokes exceeding 10 kA. Similar comparisons with local precision network lightning data from various regions (such as Sao Paulo, Brazil during the recent CHUVA campaign) are also presented. Examples of global dataset patterns and individual storms are shown. Ongoing improvements in performance, including efforts to minimize location accuracy and peak current variability across day/night boundaries, are discussed.