BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Adaptive stochastic trajectory modelling of transport in geophysic
 al flows - Esler\, JG (University College London)
DTSTART:20131203T140000Z
DTEND:20131203T144500Z
UID:TALK49149@talks.cam.ac.uk
CONTACT:Mustapha Amrani
DESCRIPTION:Motivated by the goal of improving and augmenting stochastic L
 agrangian models of particle dispersion in turbulent geophysical flows\, t
 echniques from the theory of stochastic processes are applied to a model t
 ransport problem. The aim is to find an efficient and accurate method to c
 alculate the total tracer transport between a source and a receptor when t
 he flow between the two locations is weak\, rendering direct stochastic La
 grangian simulation prohibitively expensive. Two methods are found to be u
 seful. The first is Milstein's `measure transformation method'\, which inv
 olves adding an artificial velocity to the trajectory equation\, and simul
 taneously correcting for the weighting given to each particle under the ne
 w flow. Various difficulties associated with making an appropriate choice 
 for the artificial velocity field are detailed and addressed. The second m
 ethod is a variant of Grassberger's `go-with-the-winners' branching proces
 s\, which acts to remove particles unlikely to contribute to the net trans
 port\, and reproduces those that will contribute. A simple solution to the
  problem of defining a `winner' for flows in a high Peclet number chaotic 
 advection regime is proposed. It is demonstrated that\, used independently
  or together\, the two methods can act to reduce the variance of estimator
 s of the total transport by several orders of magnitude compared with dire
 ct simulation.\n
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR