Use of state-space modelling to identify ecological covariates associated with trends in pinniped demography

Marjolaine Caillat (Corresponding Author), Line Cordes, Paul Thompson, Jason Matthiopoulos, Sophie Smout

Research output: Contribution to journalArticle

Abstract

1. Identifying and understanding ecological drivers that influence wildlife populations is challenging but critical for conservation. This typically requires integrating long-term data on both the population and potential drivers within statistical models that are suitable for analysing these complex relationships. State-space models offer one method for integrating such data. Once implemented within a Bayesian framework, these analyses can control for multifactorial influences on populations, allowing one to extract otherwise undetectable correlations between the environment and the underlying, inferred demography.

2. In the Moray Firth, Scotland, harbour seals have been counted annually for 30 years (1988-2018). A Bayesian state-space model, was used to explore whether patterns in vital rates were correlated to changes in prey abundance, inter-specific competition (grey seal abundance), environmental variables (NAO and SST) or level of biotoxins (saxitoxin and domoic acid) in the Moray Firth waters.

3. The credible interval of the posterior distributions of three of these covariate coefficients (sandeel proxy, NAO, and grey seal abundance) suggested that there was a relationship between those covariates and vital rates. Both the sandeel proxy and NAO showed a positive correlation with fecundity, whereas grey seal abundance had a negative impact on pup survival.

4. This work demonstrates how an integrated state-space modelling approach can bring together diverse data sets and point to important interactions with prey, and with other predators in the system. This suggests that the wider-scale management of UK harbour seal populations with their contrasting temporal trends needs to account for variation in the marine ecosystem at appropriate spatial scales, in line with current policy concerning spatial planning in the marine environment.
Original languageEnglish
Pages (from-to)101-118
Number of pages18
JournalAquatic Conservation: Marine and Freshwater Ecosystems
Volume29
Issue numberS1
Early online date6 Sep 2019
DOIs
Publication statusPublished - Sep 2019

Fingerprint

pinniped
demography
seals
Muraenidae
Phoca vitulina
modeling
harbor
saxitoxins
domoic acid
spatial planning
interspecific competition
statistical models
marine ecosystem
marine environment
pups
Scotland
fecundity
wildlife
sea surface temperature
planning

Keywords

  • Bayesian
  • environmental change
  • fecundity
  • harbour seal
  • inter-specific competition
  • population change
  • prey availability
  • survival
  • KITTIWAKE RISSA-TRIDACTYLA
  • INSHORE WATERS
  • MORAY FIRTH
  • SEASONAL-VARIATION
  • SEALS PHOCA-VITULINA
  • BREEDING SUCCESS
  • HARBOR SEALS
  • POPULATION-DYNAMICS
  • KILLER WHALES
  • NORTH-SEA

ASJC Scopus subject areas

  • Aquatic Science
  • Nature and Landscape Conservation
  • Ecology

Cite this

Use of state-space modelling to identify ecological covariates associated with trends in pinniped demography. / Caillat, Marjolaine (Corresponding Author); Cordes, Line; Thompson, Paul; Matthiopoulos, Jason; Smout, Sophie.

In: Aquatic Conservation: Marine and Freshwater Ecosystems, Vol. 29, No. S1, 09.2019, p. 101-118.

Research output: Contribution to journalArticle

Caillat, Marjolaine ; Cordes, Line ; Thompson, Paul ; Matthiopoulos, Jason ; Smout, Sophie. / Use of state-space modelling to identify ecological covariates associated with trends in pinniped demography. In: Aquatic Conservation: Marine and Freshwater Ecosystems. 2019 ; Vol. 29, No. S1. pp. 101-118.
@article{b9257b993602414398d2d7eec1f602bc,
title = "Use of state-space modelling to identify ecological covariates associated with trends in pinniped demography",
abstract = "1. Identifying and understanding ecological drivers that influence wildlife populations is challenging but critical for conservation. This typically requires integrating long-term data on both the population and potential drivers within statistical models that are suitable for analysing these complex relationships. State-space models offer one method for integrating such data. Once implemented within a Bayesian framework, these analyses can control for multifactorial influences on populations, allowing one to extract otherwise undetectable correlations between the environment and the underlying, inferred demography.2. In the Moray Firth, Scotland, harbour seals have been counted annually for 30 years (1988-2018). A Bayesian state-space model, was used to explore whether patterns in vital rates were correlated to changes in prey abundance, inter-specific competition (grey seal abundance), environmental variables (NAO and SST) or level of biotoxins (saxitoxin and domoic acid) in the Moray Firth waters.3. The credible interval of the posterior distributions of three of these covariate coefficients (sandeel proxy, NAO, and grey seal abundance) suggested that there was a relationship between those covariates and vital rates. Both the sandeel proxy and NAO showed a positive correlation with fecundity, whereas grey seal abundance had a negative impact on pup survival.4. This work demonstrates how an integrated state-space modelling approach can bring together diverse data sets and point to important interactions with prey, and with other predators in the system. This suggests that the wider-scale management of UK harbour seal populations with their contrasting temporal trends needs to account for variation in the marine ecosystem at appropriate spatial scales, in line with current policy concerning spatial planning in the marine environment.",
keywords = "Bayesian, environmental change, fecundity, harbour seal, inter-specific competition, population change, prey availability, survival, KITTIWAKE RISSA-TRIDACTYLA, INSHORE WATERS, MORAY FIRTH, SEASONAL-VARIATION, SEALS PHOCA-VITULINA, BREEDING SUCCESS, HARBOR SEALS, POPULATION-DYNAMICS, KILLER WHALES, NORTH-SEA",
author = "Marjolaine Caillat and Line Cordes and Paul Thompson and Jason Matthiopoulos and Sophie Smout",
note = "ACKNOWLEDGEMENTS We thank all those who shared data with us: Bob Swan for the kittiwake data, Eileen Bresnan for the biotoxin data and National Oceanic and Atmospheric Administration and the National Center for Atmospheric Research for SST and NAO data respectively. We thank Dave Thompson, Callan Duck and Chris Morris for their help in understanding and organizing survey data, and Marine Scotland for funding this work. Funding Information Marine Scotland. Grant Number: MMSS/001/11",
year = "2019",
month = "9",
doi = "10.1002/aqc.3130",
language = "English",
volume = "29",
pages = "101--118",
journal = "Aquatic Conservation: Marine and Freshwater Ecosystems",
issn = "1052-7613",
publisher = "Wiley-Blackwell",
number = "S1",

}

TY - JOUR

T1 - Use of state-space modelling to identify ecological covariates associated with trends in pinniped demography

AU - Caillat, Marjolaine

AU - Cordes, Line

AU - Thompson, Paul

AU - Matthiopoulos, Jason

AU - Smout, Sophie

N1 - ACKNOWLEDGEMENTS We thank all those who shared data with us: Bob Swan for the kittiwake data, Eileen Bresnan for the biotoxin data and National Oceanic and Atmospheric Administration and the National Center for Atmospheric Research for SST and NAO data respectively. We thank Dave Thompson, Callan Duck and Chris Morris for their help in understanding and organizing survey data, and Marine Scotland for funding this work. Funding Information Marine Scotland. Grant Number: MMSS/001/11

PY - 2019/9

Y1 - 2019/9

N2 - 1. Identifying and understanding ecological drivers that influence wildlife populations is challenging but critical for conservation. This typically requires integrating long-term data on both the population and potential drivers within statistical models that are suitable for analysing these complex relationships. State-space models offer one method for integrating such data. Once implemented within a Bayesian framework, these analyses can control for multifactorial influences on populations, allowing one to extract otherwise undetectable correlations between the environment and the underlying, inferred demography.2. In the Moray Firth, Scotland, harbour seals have been counted annually for 30 years (1988-2018). A Bayesian state-space model, was used to explore whether patterns in vital rates were correlated to changes in prey abundance, inter-specific competition (grey seal abundance), environmental variables (NAO and SST) or level of biotoxins (saxitoxin and domoic acid) in the Moray Firth waters.3. The credible interval of the posterior distributions of three of these covariate coefficients (sandeel proxy, NAO, and grey seal abundance) suggested that there was a relationship between those covariates and vital rates. Both the sandeel proxy and NAO showed a positive correlation with fecundity, whereas grey seal abundance had a negative impact on pup survival.4. This work demonstrates how an integrated state-space modelling approach can bring together diverse data sets and point to important interactions with prey, and with other predators in the system. This suggests that the wider-scale management of UK harbour seal populations with their contrasting temporal trends needs to account for variation in the marine ecosystem at appropriate spatial scales, in line with current policy concerning spatial planning in the marine environment.

AB - 1. Identifying and understanding ecological drivers that influence wildlife populations is challenging but critical for conservation. This typically requires integrating long-term data on both the population and potential drivers within statistical models that are suitable for analysing these complex relationships. State-space models offer one method for integrating such data. Once implemented within a Bayesian framework, these analyses can control for multifactorial influences on populations, allowing one to extract otherwise undetectable correlations between the environment and the underlying, inferred demography.2. In the Moray Firth, Scotland, harbour seals have been counted annually for 30 years (1988-2018). A Bayesian state-space model, was used to explore whether patterns in vital rates were correlated to changes in prey abundance, inter-specific competition (grey seal abundance), environmental variables (NAO and SST) or level of biotoxins (saxitoxin and domoic acid) in the Moray Firth waters.3. The credible interval of the posterior distributions of three of these covariate coefficients (sandeel proxy, NAO, and grey seal abundance) suggested that there was a relationship between those covariates and vital rates. Both the sandeel proxy and NAO showed a positive correlation with fecundity, whereas grey seal abundance had a negative impact on pup survival.4. This work demonstrates how an integrated state-space modelling approach can bring together diverse data sets and point to important interactions with prey, and with other predators in the system. This suggests that the wider-scale management of UK harbour seal populations with their contrasting temporal trends needs to account for variation in the marine ecosystem at appropriate spatial scales, in line with current policy concerning spatial planning in the marine environment.

KW - Bayesian

KW - environmental change

KW - fecundity

KW - harbour seal

KW - inter-specific competition

KW - population change

KW - prey availability

KW - survival

KW - KITTIWAKE RISSA-TRIDACTYLA

KW - INSHORE WATERS

KW - MORAY FIRTH

KW - SEASONAL-VARIATION

KW - SEALS PHOCA-VITULINA

KW - BREEDING SUCCESS

KW - HARBOR SEALS

KW - POPULATION-DYNAMICS

KW - KILLER WHALES

KW - NORTH-SEA

UR - http://www.scopus.com/inward/record.url?scp=85071891869&partnerID=8YFLogxK

UR - http://www.mendeley.com/research/statespace-modelling-identify-ecological-covariates-associated-trends-pinniped-demography

U2 - 10.1002/aqc.3130

DO - 10.1002/aqc.3130

M3 - Article

VL - 29

SP - 101

EP - 118

JO - Aquatic Conservation: Marine and Freshwater Ecosystems

JF - Aquatic Conservation: Marine and Freshwater Ecosystems

SN - 1052-7613

IS - S1

ER -