Data Assimilation Meetings at Reading

Date Meeting type Speakers
6 November 2013 Internal Speaker Elske van der Vaart (University of Reading)
ABC for IBMs: Parametrising Earthworms
Individual-based models, or IBMs, are becoming increasingly popular within ecology. IBMs are computer simulations that explicitly simulate individual organisms interacting with each other and their surroundings, as defined by their own characteristics and decision rules. They are a powerful tool for understanding how populations can evolve across time and space, and they are used for both fundamental research and practical ecosystem management. However, due to their complexity, they are difficult to systematically parametrise and evaluate. One promising way forward is to use Approximate Bayesian Computation, or ABC, for this purpose. Originally developed to solve problems within population genetics, ABC is a technique for approximating the posterior distributions of parameter values given available observations, without having to compute exact likelihoods. In this talk, we describe our initial effort to use ABC to parametrise an existing IBM of earthworms, as they grow and reproduce in different laboratory settings. We compare how well the parameter values found by ABC fit the empirical data relative to those chosen by hand, and we investigate the shape of resulting posterior distributions. We conclude with some ideas for future directions.

Page navigation