VOLCANO: IAVCEI 2013 session: Testing eruption forecasts from statistical, geophysical, and laboratory models

[Jean-François Smekens <jsmekens@xxxxxxx>]

****************************************************************************************
IAVCEI 2013 session: Testing eruption forecasts from statistical, geophysical, and laboratory models
From: Andrew Bell <a.bell@xxxxxxxx>
****************************************************************************************

Dear All,
    
    We look forward to seeing you at IAVCEI 2013, especially at the
    following
    session on testing eruption forecasts (Session 4.4).
    
    Testing eruption forecasts from statistical, geophysical, and
      laboratory models
      
    
    http://www.iavcei2013.com/scientific_program/scientific_program.html#4_4
    
    
    How well can eruptions be forecast? This question can be answered
    rigorously only when a forecast is falsifiable, i.e., when temporal
    and,
    ideally, spatial and magnitude windows are specified, along with a
    probability of an eruption occurring within those windows. Given
    these
    requirements, and a sufficiently large sample of forecasts, it is
    possible to test whether a forecasting technique possesses any
    inherent
    skill. Forecasts may be issued for temporal scales from minutes
    (e.g.,
    for the automatic shutdown of critical facilities) to centuries
    (e.g.,
    for optimizing land-management policies). Whatever the time scale,
    successful forecasts, false alarms, and eruptions that are not
    forecast,
    all incur a cost to lives and infrastructure, from the consequences
    of
    emergency procedures, to long-term decisions on land use. It is thus
    crucial to evaluate the relative skill of current forecasting
    methods and
    to identify procedures for improving the quality of forecasts in the
    future. 
    
    We seek papers that: (1) evaluate forecasting techniques, either
    retrospectively or prospectively, where the testing data are
    independent
    of the model selection and calibration, and are not included in the
    model
    specification; (2) describe modifications to forecasting models in
    response to predictive performance; (3) evaluate the performance of
    monitoring parameters; and (4) codify the procedures for evaluating
    forecasting methods. We welcome contributions based on catalog data,
    geophysical field observations and monitoring data, laboratory
    experiments, and numerical modelling of precursory phenomena.
    
    The abstract deadline is 31st Jan 2013:
      
    
    http://www.iavcei2013.com/abstract_submission/abstract_submission.html
    
    
    Best wishes,
    
    Andrew Bell, Mark Bebbington, Gill Jolly and Christopher Kilburn
  



==============================================================

Volcano Listserv is a collaborative venture among Arizona State University (ASU), Portland State University (PSU), the Global Volcanism Program (GVP) of the Smithsonian Institution's National Museum of Natural History, and the International Association for Volcanology and Chemistry of the Earth's Interior (IAVCEI).

ASU - http://www.asu.edu/ PSU - http://pdx.edu/ GVP - http://www.volcano.si.edu/ IAVCEI - http://www.iavcei.org/

To unsubscribe from the volcano list, send the message: signoff volcano to: listserv@xxxxxxx, or write to: volcano-request@xxxxxxx.

To contribute to the volcano list, send your message to: volcano@xxxxxxx. Please do not send attachments.

==============================================================