BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:EESS talk on "Searching for an attractor in large radar data archi
 ves  to study the predictability of precipitation"
DTSTART:20160920T121500
DTEND:20160920T131500
DTSTAMP:20260531T011240Z
UID:fb3ca1a04ce42b32890c0b138c668ecd099e2b86a9ee12b6cd265511
CATEGORIES:Conferences - Seminars
DESCRIPTION:Dr Loris Foresti\, MeteoSwiss\, Locarno-Monti\nAbstract: Very 
 short-term precipitation forecasting (0-6 hours)\, usually known as nowcas
 ting\, is a fundamental ingredient in early warning systems for urban floo
 ds\, landslides\, severe thunderstorms and hail. The fractal behaviour and
  high variability of precipitation in space and time makes both its measur
 ement and prediction particularly challenging. As a consequence\, it is pr
 actically impossible to provide an accurate deterministic quantitative pre
 cipitation forecast and more efforts should be spent in estimating the for
 ecast uncertainty.\n\nUsing as basis the concept of analogues\, which firs
 t appeared in the idealized Lorenz attractor\, the core idea of this Ambiz
 ione project is to present a new framework to construct a strange attracto
 r for precipitation directly from the large radar and satellite data archi
 ves. The ensemble of analogue precipitation patterns retrieved from the at
 tractor is expected to give new insights into the intrinsic predictability
  of precipitation and could be exploited to design adaptive nowcasting sys
 tems that better represent the forecast uncertainty.\n\nShort biography: L
 oris Foresti is research scientist at MeteoSwiss and is leading a Swiss Na
 tional Science Foundation (SNSF) Ambizione project on the short-term predi
 ctability of precipitation using the large radar and satellite data archiv
 es. Loris Foresti was born in Locarno\, Switzerland\, in 1985\, and studie
 d environmental geosciences at the University of Lausanne\, where he also 
 received a PhD degree in 2011. The topic of the thesis was the spatial int
 erpolation of meteorological variables in complex orography using kernel-b
 ased machine learning methods. After obtaining the PhD he has been 4 years
  abroad working as postdoctoral scientist at the Australian Bureau of Mete
 orology\, Melbourne (supported by SNSF) and at the Royal Meteorological In
 stitute of Belgium. In both places he carried out cutting edge research on
  the predictability of precipitation and also developed a real-time probab
 ilistic precipitation nowcasting system based on the extrapolation of rada
 r images in Belgium.
LOCATION:GR A3 31 http://plan.epfl.ch/?room=GR%20A3%2031
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR