layoutGraphic

Workshops

Workshops 2010

  • Introduction to Distance Sampling

    24th - 27th August 2010, St Andrews

  • Advanced Techniques and Recent Developments in Distance Sampling

    • 30th August - 1st September 2010, St Andrews

Past Workshops

7th - 10th September 2009, St Andrews

24th - 28th August 2009

18th - 21st August 2009

3rd - 7th August 2009

 

Conferences

International Statistical Ecology Conference 6 - 9 July 2010 (ISEC 2010)

International Statistical Ecology Conference 9 - 11 July 2008 (ISEC)

Seminars

NCSE seminars are transmitted by video conference to all three sites, and are often broadcast by a remote speaker from a fourth site. All are welcome to attend.

    Current Seminars

  • 14 January 2010 16.00 UK time
    Perry de Valpine, University of California Davis, USA.
    Classical analysis of state-space models for population dynamics.

    • State-space models of population dynamics combine two stochastic models: one for the true population dynamics and another for data sampled from the true population. The difficulty of working with state-space models is that the probability of the data given some parameters, namely the likelihood, requires summation over all of the possible true state trajectories from which the data might have been observed. In even simple models with density-dependence or age-structure, one often has a nonlinear and/or non-Gaussian model, so the Kalman Filter does not apply. I will review approaches to classical maximum likelihood estimation of these nonlinear and/or non-Gaussian state-space models, including numerical filtering methods and methods based on Markov chain Monte Carlo. The latter include Monte Carlo Expectation Maximization, data cloning, and Monte Carlo Kernel Likelihood, for which I will give accuracy and efficiency results. I will briefly review the problem of estimating normalizing constants so that normalized likelihoods can be used in AIC or likelihood ratio comparisons. The bridge sampling approach to normalizing constant estimation will be emphasized. Applicability of these methods to integrated population modeling, in which a state-space model and capture-mark-recapture or ring-recovery model are combined, will be illustrated with the British Lapwing data. Finally I will introduce initial progress on estimation of nonparametric state-space models.

    Past Seminars

  • 15 December 2008
    Gary White, Colorado State University, USA.
    Program MARK Overview and Recent Additions.
  • 27 November 2007
    Geir Storvik, University of Oslo, Norway.
    Modelling Pollock egg counts from the western Gulf of Alaska by a zero-inflated Bayesian hierarchical space-time model
  • 14 November 2007
    Ed Ionides, University of Michigan, USA.
    Inference for nonlinear dynamical systems, with applications to the ecology of infectious diseases. (slides)
  • 19 October 2007
    Jon Barry, CEFAS, UK.
    How many benthic species are there and how bad is dredging for them?
  • 24 May 2007
    Jean-Michel Gaillard, University of Lyon, France.
    How does individual heterogeneity influence detection of senescence and trade-offs: ungulates as case studies.
  • 16 May 2007
    Paul Conn, Colorado State University, USA.
    Bayesian analysis of wildlife age-at-harvest data
  • 5 February 2007
    William Browne, University of Nottingham, UK.
    Using complex random effect models in epidemiology and ecology
  • 29 November 2006
    Carmen Fernandez, Spanish Oceanographic Institute.
    Inference for state space models of wild animal populations
  • 30 August 2006
    David Fletcher, University of Otago, Dunedin, New Zealand.
    Mark-Recapture Models and Population Dynamics
  • 19 June 2006
    Mark Maunder and Rick Deriso, Inter American Tropical Tuna Commission, San Diego, USA.
    Including covariates in population dynamics models
  • 12 June 2006
    Dave Thomson, Max Planck Institute for Demographic Research, Rostock, Germany.
    Statistical Analyses in Biodemography