Harvest‐based Bayesian estimation of sika deer populations using state‐space models
- 21 December 2007
- journal article
- Published by Wiley in Population Ecology
- Vol. 50 (2) , 131-144
Abstract
We have estimated the number of sika deer, Cervus nippon, in Hokkaido, Japan, with the aim of developing a management program that will reduce the level of agricultural damage caused by these deer. A population index that is defined by the population divided by the population of 1993 is first estimated from the data obtained during a spotlight survey. A generalized linear mixed model (GLMM) with corner point constraints is used in this estimation. We then estimate the population from the index by evaluating the response of index to the known amount of harvest, including hunting. A stage‐structured model is used in this harvest‐based estimation. It is well‐known that estimates of indices suffer from large observation errors when the probability of the observation fluctuates widely; therefore, we apply state‐space modeling to the harvest‐based estimation to remove the observation errors. We propose the use of Bayesian estimation with uniform prior‐distributions as an approximation of the maximum likelihood estimation, without permitting an arbitrary assumption that the parameters fluctuate following prior‐distributions. We are able to demonstrate that the harvest‐based Bayesian estimation is effective in reducing the observation errors in sika deer populations, but the stage‐structured model requires many demographic parameters to be known prior to running the analyses. These parameters cannot be estimated from the observed time‐series of the index if there is insufficient data. We then construct a univariate model by simplifying the stage‐structured model and show that the simplified model yields estimates that are nearly identical to those obtained from the stage‐structured model. This simplification of the model simultaneously clarifies which parameter is important in estimating the population.Keywords
This publication has 45 references indexed in Scilit:
- How to analyze long‐term insect population dynamics under climate change: 50‐year data of three insect pests in paddy fieldsPopulation Ecology, 2005
- State-space likelihoods for nonlinear fisheries time-seriesCanadian Journal of Fisheries and Aquatic Sciences, 2005
- Mechanisms of density dependence in fluctuating vole populations: deducing annual density dependence from seasonal processesPopulation Ecology, 2003
- BETTER INFERENCES FROM POPULATION-DYNAMICS EXPERIMENTS USING MONTE CARLO STATE-SPACE LIKELIHOOD METHODSEcology, 2003
- POPULATION DYNAMICS ACROSS GEOGRAPHICAL RANGES: TIME-SERIES ANALYSES OF THREE SMALL GAME SPECIESEcology, 2003
- Extending production models to include process error in the population dynamicsCanadian Journal of Fisheries and Aquatic Sciences, 2003
- Modeling temporal and spatial colony‐site dynamics in a long‐lived seabirdPopulation Ecology, 2003
- Confusion Over Measures of Evidence (p's) Versus Errors (α's) in Classical Statistical TestingThe American Statistician, 2003
- Bayesian Measures of Model Complexity and FitJournal of the Royal Statistical Society Series B: Statistical Methodology, 2002
- Markov Chain Monte Carlo Convergence Diagnostics: A Comparative ReviewJournal of the American Statistical Association, 1996