Details:

Journal Title:
Fisheries Research
Personal Author:
Charsley, A.R. ; Grüss, A. ; Thorson, J.T. ; Rudd, M.B. ; Crow, S.K. ; ... More +
Charsley, A.R. ; Grüss, A. ; Thorson, J.T. ; Rudd, M.B. ; Crow, S.K. ; David, B. ; Williams, E.K. ; Hoyle, S.D. Less -
NOAA Program & Office:
NMFS (National Marine Fisheries Service) ; AFSC (Alaska Fisheries Science Center)
NMFS (National Marine Fisheries Service) ; AFSC (Alaska Fisheries Science Center) Less -
Description:
Spatio-temporal modelling frameworks are important tools for evaluating changes in the population structure of freshwater species, to inform population assessments. However, in the stream networks occupied by freshwater species, two data points in space are more related by physical connectivity than by straight-line, Euclidian distance. Therefore, there is a need for spatio-temporal modelling frameworks that represent the relatedness of points based on physical connectivity. We developed such a modelling framework, known as the “VAST stream network modelling framework", using the widely-employed vector autoregressive spatio-temporal modelling platform VAST and the Ornstein-Uhlenbeck process for spatial variance. Here, we demonstrate the VAST stream network spatio-temporal modelling platform with two applications to the freshwater stages of a diadromous fish species, the endemic New Zealand longfin eel (Anguilla dieffenbachii), and a simulation experiment. The first real-world application considers a single category (longfin eel) and models its probability of encounter in the Waitaki catchment, while the second distinguishes between two categories (smaller and larger longfin eels) and models their numerical densities in the greater Waikato region. These two applications demonstrate how stream network spatio-temporal models can usefully inform freshwater resource managers, providing insights into changes in fish encounter probability and density for different fish length classes and estimates of effective river length occupied. The simulation experiment uses the Waikato model as an operating model to evaluate alternative sampling scenarios for accuracy, precision and coverage, to advise on data requirements. A yearly sample size of 30 sites over a 20-year time series had the greatest precision and accuracy in trend estimates among all scenarios investigated. We also found that, for the scenario with declining female spawning biomass, increasing the yearly sample size from 15 sites to 30 sites in a 9-year time series increased the precision and resulted in unbiased estimates. There is a need to optimise sampling schemes for the freshwater stages of diadromous species, including longfin eel, and we, therefore, recommend further simulation experiments to explore scenarios of population trends, habitat features and sampling approaches.
Keywords:
[+]
Anguilla Dieffenbachii Diadromous Fishes Ornstein-Uhlenbeck Process Diadromous Species Longfin Eel Spatio-temporal Stream Network Models VAST Modelling Platform
Source:
Fisheries Research 259 (2023) 106583
DOI:
https://doi.org/10.1016/j.fishres.2022.106583
Format:
PDF
Document Type:
Journal Article
Rights Information:
Accepted Manuscript
Compliance:
Submitted
Main Document Checksum:
[+]
urn:sha256:b3f61d4b1096039182617d0ce0610a320583e6d01a576f452ee9d186fe7ffdc0
Download URL:
https://repository.library.noaa.gov/view/noaa/56233/noaa_56233_DS1.pdf
File Type:

Supporting Files

• Supplemental media

  docx

More +