Integrated Population Modelling
Monday 24th September 2012
|Document downloads for IBS members.|
Join us now.
|11.00 - 11.30||Registration|
|11.30 - 12.30||Tutorial talk on Integrated Population Modelling, by Takis Besbeas|
(Athens University of Business and Economics, University of Kent)
|12.30 - 13.50||Lunch|
|12.30 - 13.50||Poster Session|
Below are the posters to be presented at this session. (The presenting author is in bold):
Sheikh Taslim Ali (Imperial College London and Karnatak University, India) "Bayesian Stochastic Modelling: An approach to quantify the Transmission Dynamics of the 2009 H1N1 pandemic"
Takis Besbeas, Rachel McCrea and Byron Morgan (University of Kent) “Model selection for Integrated Population Models: Transdimensional state-vectors and survival age-structure selection.”
Diana Cole and Rachel McCrea (University of Kent). “Parameter redundancy in Integrated Population Models.”
David Conesa, Antonio López-Quílez, Silvia Lladosa, Facundo Muñoz (Universitat de València), Maria Grazia Penninob, Josè Maria Bellido (Centro Oceanográfico de Murcia), Janine Illian (University of St Andrews), Daniel Simpson (University of Helsinki), Marta González-Warleta, Mercedes Mezo (Centro de Investigaciones Agrarias de Mabegondo-INGACAL), “Assessing the spatial distribution of species using Bayesian latent Gaussian models.”
Ben Hubbard, Diana Cole and Byron Morgan (University of Kent) “Parameter redundancy: Identifiability issues in Mark-Recapture-Recovery models”
Jose Lahoz-Monfort (University of Kent), Byron Morgan (University of Kent), Mike Harris(Centre for Ecology & Hydrology), Sarah Wanless (Centre for Ecology & Hydrology) and Stephen Freeman (Centre for Ecology & Hydrology) “Bringing it all together: a multi-species integrated population model for a breeding community of seabirds”
Stephan Schiffels and Richard Durbin (Wellcome Trust Sanger Institute.) “The Multiple Sequential Markovian Coalescent (MSMC)”
Hannah Worthington and Ruth King (University of St Andrews) “Integrated Stopover Models”
|13.30 - 13.45||Environmental Statistics Section of the Royal Statistical Society AGM|
|13.50 - 14.30||Recent developments in integrated population modelling, by Takis Besbeas|
(Athens University of Business and Economics, University of Kent)
The most common type of information from a monitoring scheme is a time series of population abundances but the potential of such data for analysing population dynamics is limited. Recent work has shown that repeating the population survey can considerably increase the information obtained from the monitoring. We introduce new approaches to incorporating information from repeated surveys in state-space models for population dynamics. The methods include an approach based on the use of population estimates and standard errors, which is simpler and lends itself readily to data derived by any monitoring program. We then investigate the use of replicate sampling in Integrated Population Monitoring. We show that replication improves estimation of sampling error, as expected, but it also allows the use of more complex process error structures than usually considered. Using simulation, we evaluate its performance against the case of no replication as well as no replication but instead observing more states for age-structured populations. Based on the results, we provide some general recommendations on how sampling effort should be allocated in population surveys to improve inference.
|14.30 - 15.10||Combining demographic and population count data to estimate immigration rate and the strength of |
density dependence, by Fitsum Abadie (Centre D'Ecologie Fonctionnelle & Evolutive)
The recently developed integrated population modeling allows the use of demographic and population count data in a coherent fashion to estimate as well as model demographic parameters as a function of covariates. In this talk, I will present the potential of integrated population modeling in estimating immigration rate as well as studying the strength of density dependence. I will demostrate the application of the model using long-term data of the little owl (Athene noctua) and red-backed shrike (Lanius collurio) populations from Southern Germany.
|15.10 - 15.40||Coffee Break|
|15.40 - 16.20||State-space modeling reveals the proximate causes of harbour seal population declines|
by Jason Matthiopoulos (University of St. Andrews)
Declines in large vertebrates are widespread but difficult to detect from monitoring data and hard to understand due to a multiplicity of biological explanations. In the north and east of Scotland counts of harbour seals (Phoca vitulina) have been declining for 10 years. To evaluate the contributions of different proximate causes (survival, fecundity, observation artifacts) to this decline, we collated behavioural, demographic and population data from an intensively studied population in part of the Moray Firth (NE Scotland). To these, we fitted a state-space model comprising age-structured dynamics and a detailed account of observation errors. Our results confirm that the trends in the population counts are the result of an underlying decline in population numbers, not an artifact of the observation process. After accounting for the effect of culling, the main driver of the population decline is a decreasing trend in survival, particularly of juvenile individuals combined with (previously undetected) low historical levels of pupping success. The model provides evidence for considerable increases in breeding success and consistently high levels of adult survival, hinting that adults are unaffected by the mechanisms influencing juveniles (e.g. intraspecific competition or competition with other species such as grey seals). Forecasts from the model indicate a slow recovery in the Moray Firth, providing cautious support for the continuation of the Moray Firth Seal Management Plan.
The investigation of proximate causes (survival, fecundity and observation errors) is a quick-response mechanism for the management of population declines. It can help exclude entire categories of ultimate causal mechanisms hence focusing future field data collection. The contribution of specific ultimate drivers (e.g. shooting or competitors) can also be quantified by including them as covariates to survival or fecundity. From an applied perspective, this study of a relatively isolated and well-studied population has important implications for harbour seals in the rest of the UK, some of which have shown considerably steeper declining trends.
|16.20 - 17.00||Dynamic probabilistic models for predicting regime shifts in fish populations, by Allan Tucker |
In this talk I will discuss how different architectures of Bayesian networks have led to interesting discoveries in molecular biology before discussing the ‘crossover potential’ of such techniques to ecological domains with an example using fisheries data. I will show how dynamic Bayesian networks with latent variables can be used to predict population collapses in different ocean regions. The difficulties arising from different species performing similar functions in the different regions will be explored as will the importance of the availability of a number of independent datasets for discovering genuine interactions.
|17.00 - 17.20||Discussion led by Byron Morgan (University of Kent)|
Existing members can login below to view all site content. Lost password?
Other visitors might be interested to learn more about the benefits of membership.
|14 Nov 14||Mendelian Randomisation: Past Success and Future Challenges|
|17 Sep 14||Statistics for big data|
|06 Jul 14 - 11 Jul 14||International Biometric Conference 2014, Florence|
|04 Jun 14||Clinical trials: past, present and glorious future?|
|14 Nov 13||Statistical software interoperability|