|
| |
|
|
Possibilities of Using of Remote Detection Data for Convective Storms Intensity Nowcasting
Valachová, Michaela ; Žák, Michal (advisor) ; Bližňák, Vojtěch (referee) ; Simon, André (referee)
Title: Possibilities of Using of Remote Detection Data for Convective Storms Intensity Nowcasting Author: Michaela Valachová Department: Department of Atmospheric Physics Supervisor: Mgr. Michal Žák, Ph.D., Department of Atmospheric Physics Abstract: Evolution of 60 isolated convective storms from 2016 and 2017, which formed in the region of Central Europe, is studied by means of multi-sensor observations. According to the reports from the European Severe Weather Da- tabase, two categories of storms are classified: severe and non-severe. Based on radar, lightning and satellite measurements, trends of storm characteristics are analyzed to ascertain their typical behavior. Lightning stroke rates and their change could well warn about the ability of the storm to become severe, therefore a Lightning jump algorithm was proposed within this work. From individual case studies follows that methods of remote sensing offer comprehensive information about convective storm life-cycles. In order to objectively determine crucial variables for estimating the storm se- verity, logistic regression models and regularized regressions (elastic net) are employed. In total 53 variables from the first 30, 60 and 90 minutes of the moni- tored storm lifetime are used to show their predictive skill. Results of the models indicate...
|
|
|
Possibilities of Using of Remote Detection Data for Convective Storms Intensity Nowcasting
Valachová, Michaela ; Žák, Michal (advisor) ; Bližňák, Vojtěch (referee) ; Simon, André (referee)
Title: Possibilities of Using of Remote Detection Data for Convective Storms Intensity Nowcasting Author: Michaela Valachová Department: Department of Atmospheric Physics Supervisor: Mgr. Michal Žák, Ph.D., Department of Atmospheric Physics Abstract: Evolution of 60 isolated convective storms from 2016 and 2017, which formed in the region of Central Europe, is studied by means of multi-sensor observations. According to the reports from the European Severe Weather Da- tabase, two categories of storms are classified: severe and non-severe. Based on radar, lightning and satellite measurements, trends of storm characteristics are analyzed to ascertain their typical behavior. Lightning stroke rates and their change could well warn about the ability of the storm to become severe, therefore a Lightning jump algorithm was proposed within this work. From individual case studies follows that methods of remote sensing offer comprehensive information about convective storm life-cycles. In order to objectively determine crucial variables for estimating the storm se- verity, logistic regression models and regularized regressions (elastic net) are employed. In total 53 variables from the first 30, 60 and 90 minutes of the moni- tored storm lifetime are used to show their predictive skill. Results of the models indicate...
|
|
| |
guest ::
RSS feed.