TITLE:

Risk estimation and mapping using a bayesian approach

DATE:

Friday, December 9th, 2016

TIME:

3:30 PM

LOCATION:

GMCS 314

SPEAKER:

Dr. Lelys Bravo. Visiting Professor, Department of Applied

Mathematics and Statistics. University of California in Santa

Cruz.

ABSTRACT:

This talk presents a methodology to estimate environmental risk,

departing from its standard definition: the product of the

probability of a given hazard, times the vulnerability associated

to it. Vulnerability is estimated as a loss function. This

function quantifies different kinds of losses, including people

affected, or economic losses associated with a certain level of

the hazard. It represents the intrinsic condition of the exposed

population or infrastructure to suffer damage at a given time and

location. Risk is finally defined as the expected loss integrated

over all possible values of the hazard, and the losses associated

with each hazard level. Uncertainties in losses and hazard levels

can be incorporated by using appropiate probability models for

both variables. Under a bayesian framework, simulations from the

joint posterior predictive distribution of hazard and

vulnerability are used to evaluate the expected loss or final

risk. Space-time modeling of vulnerability and hazard, will

provide a direct output for risk mapping. Examples are presented

for risk estimation to extreme rainfall events in Venezuela.

HOST:

Dr. Jose Castillo

DOWNLOAD: