A hypothesis-driven statistical approach for identifying ecosystem indicators of coho and Chinook salmon marine survival

Efforts to understand causes of declines in productivity of species of concern often involve retrospective evaluation of multiple possible causes based on trends in relevant ecological indicators. We describe a hypothesis testing framework for examining declines in marine survival for coho and Chino...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Kathryn L. Sobocinski, Correigh M. Greene, Joseph H. Anderson, Neala W. Kendall, Michael W. Schmidt, Mara S. Zimmerman, Iris M. Kemp, Su Kim, Casey P. Ruff
Formato:	article
Lenguaje:	EN
Publicado:	Elsevier 2021
Materias:	Salmon Marine survival Ecosystem Indicators Ocean GAM Ecology QH540-549.5
Acceso en línea:	https://doaj.org/article/2107eb3fd23743aa8f355b3affd15b31
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:2107eb3fd23743aa8f355b3affd15b31
record_format	dspace
spelling	oai:doaj.org-article:2107eb3fd23743aa8f355b3affd15b312021-12-01T04:45:30ZA hypothesis-driven statistical approach for identifying ecosystem indicators of coho and Chinook salmon marine survival1470-160X10.1016/j.ecolind.2021.107403https://doaj.org/article/2107eb3fd23743aa8f355b3affd15b312021-05-01T00:00:00Zhttp://www.sciencedirect.com/science/article/pii/S1470160X21000686https://doaj.org/toc/1470-160XEfforts to understand causes of declines in productivity of species of concern often involve retrospective evaluation of multiple possible causes based on trends in relevant ecological indicators. We describe a hypothesis testing framework for examining declines in marine survival for coho and Chinook salmon in the Salish Sea. Independent populations of both anadromous species have declined over the last 50 years, prompting extensive examination of mortality in different life stages. Previous studies have identified declining trends in marine survival, and we re-evaluated these trends in light of a number of possible hypotheses for declines. We laid out seven potential explanations for declines: changes in predator buffering related to abundance and timing, density-dependent or -independent food availability, water quality, timing of freshwater delivery to Puget Sound, and anthropogenic impacts. We compiled ecosystem indicators relevant to these hypotheses and used generalized additive models (GAMs) to examine multivariate relationships with survival from multiple coho and Chinook salmon stocks. We also developed additional models using the most informative indicators based on variable importance weighting (VIW) from the seven hypothesis groups. We examined how these models explained overall trends in marine survival, as well as survival in three temporal stanzas (before, during, and after a major decline, based on statistical breakpoint analysis). Across the entire time series, best fitting models explained 30–40% of the variation in the survival data. Best fitting models were from multiple hypotheses, including predation (abundance and timing), competition, water quality, and anthropogenic impacts; the freshwater delivery hypothesis was the least supported. Different models performed best (lowest error) during different stanzas of the coho salmon marine survival time series and the two VIW models were generally the top performing models, but performance varied in different years. Indicators with the strongest support included seal abundance, herring abundance, timing of hatchery salmon releases, and indicators related to water properties like stratification and temperature. These findings suggest that multiple processes embedded in several of our hypotheses influence marine survival but that an ecological “smoking gun” for Salish Sea salmon declines will remain elusive.Kathryn L. SobocinskiCorreigh M. GreeneJoseph H. AndersonNeala W. KendallMichael W. SchmidtMara S. ZimmermanIris M. KempSu KimCasey P. RuffElsevierarticleSalmonMarine survivalEcosystemIndicatorsOceanGAMEcologyQH540-549.5ENEcological Indicators, Vol 124, Iss , Pp 107403- (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Salmon Marine survival Ecosystem Indicators Ocean GAM Ecology QH540-549.5
spellingShingle	Salmon Marine survival Ecosystem Indicators Ocean GAM Ecology QH540-549.5 Kathryn L. Sobocinski Correigh M. Greene Joseph H. Anderson Neala W. Kendall Michael W. Schmidt Mara S. Zimmerman Iris M. Kemp Su Kim Casey P. Ruff A hypothesis-driven statistical approach for identifying ecosystem indicators of coho and Chinook salmon marine survival
description	Efforts to understand causes of declines in productivity of species of concern often involve retrospective evaluation of multiple possible causes based on trends in relevant ecological indicators. We describe a hypothesis testing framework for examining declines in marine survival for coho and Chinook salmon in the Salish Sea. Independent populations of both anadromous species have declined over the last 50 years, prompting extensive examination of mortality in different life stages. Previous studies have identified declining trends in marine survival, and we re-evaluated these trends in light of a number of possible hypotheses for declines. We laid out seven potential explanations for declines: changes in predator buffering related to abundance and timing, density-dependent or -independent food availability, water quality, timing of freshwater delivery to Puget Sound, and anthropogenic impacts. We compiled ecosystem indicators relevant to these hypotheses and used generalized additive models (GAMs) to examine multivariate relationships with survival from multiple coho and Chinook salmon stocks. We also developed additional models using the most informative indicators based on variable importance weighting (VIW) from the seven hypothesis groups. We examined how these models explained overall trends in marine survival, as well as survival in three temporal stanzas (before, during, and after a major decline, based on statistical breakpoint analysis). Across the entire time series, best fitting models explained 30–40% of the variation in the survival data. Best fitting models were from multiple hypotheses, including predation (abundance and timing), competition, water quality, and anthropogenic impacts; the freshwater delivery hypothesis was the least supported. Different models performed best (lowest error) during different stanzas of the coho salmon marine survival time series and the two VIW models were generally the top performing models, but performance varied in different years. Indicators with the strongest support included seal abundance, herring abundance, timing of hatchery salmon releases, and indicators related to water properties like stratification and temperature. These findings suggest that multiple processes embedded in several of our hypotheses influence marine survival but that an ecological “smoking gun” for Salish Sea salmon declines will remain elusive.
format	article
author	Kathryn L. Sobocinski Correigh M. Greene Joseph H. Anderson Neala W. Kendall Michael W. Schmidt Mara S. Zimmerman Iris M. Kemp Su Kim Casey P. Ruff
author_facet	Kathryn L. Sobocinski Correigh M. Greene Joseph H. Anderson Neala W. Kendall Michael W. Schmidt Mara S. Zimmerman Iris M. Kemp Su Kim Casey P. Ruff
author_sort	Kathryn L. Sobocinski
title	A hypothesis-driven statistical approach for identifying ecosystem indicators of coho and Chinook salmon marine survival
title_short	A hypothesis-driven statistical approach for identifying ecosystem indicators of coho and Chinook salmon marine survival
title_full	A hypothesis-driven statistical approach for identifying ecosystem indicators of coho and Chinook salmon marine survival
title_fullStr	A hypothesis-driven statistical approach for identifying ecosystem indicators of coho and Chinook salmon marine survival
title_full_unstemmed	A hypothesis-driven statistical approach for identifying ecosystem indicators of coho and Chinook salmon marine survival
title_sort	hypothesis-driven statistical approach for identifying ecosystem indicators of coho and chinook salmon marine survival
publisher	Elsevier
publishDate	2021
url	https://doaj.org/article/2107eb3fd23743aa8f355b3affd15b31
work_keys_str_mv	AT kathrynlsobocinski ahypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival AT correighmgreene ahypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival AT josephhanderson ahypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival AT nealawkendall ahypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival AT michaelwschmidt ahypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival AT maraszimmerman ahypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival AT irismkemp ahypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival AT sukim ahypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival AT caseypruff ahypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival AT kathrynlsobocinski hypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival AT correighmgreene hypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival AT josephhanderson hypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival AT nealawkendall hypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival AT michaelwschmidt hypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival AT maraszimmerman hypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival AT irismkemp hypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival AT sukim hypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival AT caseypruff hypothesisdrivenstatisticalapproachforidentifyingecosystemindicatorsofcohoandchinooksalmonmarinesurvival
_version_	1718405768199274496

A hypothesis-driven statistical approach for identifying ecosystem indicators of coho and Chinook salmon marine survival

Ejemplares similares