Publications

Shikhov, A; Chernokulsky, A (2018). A satellite-derived climatology of unreported tornadoes in forested regions of northeast Europe. REMOTE SENSING OF ENVIRONMENT, 204, 553-567.

Abstract
This study presents a novel method of tornado track identification in forested regions in Europe based on remote sensing data. The method enables an objective estimate (i.e. independent of population density and observational networks) of tornado climatology in forested regions. The method is based on the identification of narrow and elongated areas as forest disturbances obtained using Landsat satellite images and Landsat-based Global Forest Change (GFC) data. These areas were subsequently verified with high-resolution satellite images for verification of a tomadic cause of forest damage. Landsat and MODIS satellite images, weather station observations and reanalysis data were additionally involved in order to determine tornado dates. A minimum F scale tornado intensity was estimated by a Weibull distribution model using information on tornado path lengths and widths. The method is applied to the forested regions of northeast Europe, where 110 tornado tracks were identified between the 2000 and 2014 years, 105 of which were previously unreported and discovered for the first time. For some regions, tornado density estimates using the new method is 2-3 times higher than other previously published estimates. The largest number of tornadoes occurred in 2009, and June is the most favourable month for tornado formation (including strong tornadoes and tornado outbreaks). Most identified tornadoes have path length < 10 km with maximum and mean widths of approximately 200-300 m and 100-200 m, respectively. A few tornadoes with long and wide paths were found; four of them likely had F3 minimal intensity.

DOI:
10.1016/j.rse.2017.10.002

ISSN:
0034-4257