Inference for nonlinear state-space models of animal population dynamics given biased sequential life stage data with application to an endangered fish species
Speaker: Ken Newman, BioSS
State-space models (SSMs) are a popular tool for modeling animal abundances. Inference difficulties for simple linear SSMs are well known, particularly in relation to simultaneous estimation of process and observation variances. For the particular case of nonlinear stage-structured SSMs fit with biased sequential life stage data, we show theoretically, using the methods of Cole and McCrea (2016), and by simulation how covariates in models for the state process and externally estimated observation variances can potentially alleviate identifiability problems. This work was motivated by the aim to understand the effects of environmental and anthropogenic factors on the population dynamics of Delta Smelt, an endangered minnow-sized fish endemic to the San Francisco Estuary and the centre of much political and legal controversy related to water management. To increase the utility of long-term fish monitoring data for fitting SSMs, gear efficiency studies were implemented that allowed estimation of observation variances external to SSM fitting and thus improve inferences.