Movement Probabilities of Bigheaded Carps (Hypophthalmichthys spp.) in the Illinois River Estimated from Markov Chain Monte Carlo Methods

Previously published work used program MARK to estimate parameters for multistate models for Bighead Carp (Hypophthalmichthys nobilis) and Silver Carp (H. molitrix) (Coulter et al. 2018) movements in the Illinois River. Multistate models estimate transition probability (psi), survival probability (S), and detection probability (p) and models with those parameters varying by state and time (state*time), just state (state), just time (time), and neither (.) were evaluated for each species (Coulter et al. 2018). Using the format of the top model for each species identified in Coulter et al. (2018), Markov Chain Monte Carlo (MCMC) methods were used to provide a hyperdistribution of estimates for all model parameters. The top model for Bighead Carp was S(.)p(state*time)psi(state) and for Silver Carp was S(.)p(state*time)psi(state). To generate hyperdistributions for parameter estimates from MCMC chains, three chains were run for each parameter with a seed value of 0, 4,000 tuning sampling, 1,000 burn in samples, and then run for 10,000 steps. Diagnostics (provided in datasets) indicated chain convergence. Parameters with estimated values near zero from frequentist parameter estimates (explained in Coulter et al. 2018) were fixed at 0 during MCMC estimation to reduce run times. Parameter estimates represent monthly values. There are two files in this dataset: 1) SCMCMCChainsfixed.xls contains parameter estimates, MCMC chain results, and convergence metrics for Silver Carp 2) BCMCMCChainsfixed.xls contains parameter estimates, MCMC chain results, and convergence metrics for Bighead Carp File type: .zipFile size: 31 M