Ecosystem correlations with survival: stoplight modeling, single and multivariate analyses, and other approaches

Investigating Agencies: NOAA Fisheries, Department of Fisheries and Oceans, Washington Department of Fish and Wildlife, Puget Sound Treaty Tribes, U. of Washington

Ecosystem indicators are being developed and analyzed for their ability to predict the marine survival of salmon and steelhead. The objectives of the indicators work are to provide a central location for organizing and compiling metrics for the project, to determine whether the indicators can be used to improve forecasts of adult returns, and to look back through time to evaluate indicators that may have correlated with the decline in survival of Chinook, coho, and steelhead.

Similar to the survival datasets, this work was initially performed retrospectively but the datasets are updated over the course of the project where appropriate and expanded to include new data collected via the proposed sampling program. A broad suite of indicators has been established for the entire Salish Sea. This indicators list functions as the guidance document for compiling metrics used throughout the project: for indicators/correlative analyses, ecosystem modeling, and bottom-up data collection.

A stoplight modeling approach will be used to coarsely evaluate indicators across the Salish Sea basin and to ensure cross-talk between U.S. and Canada. However, finer-scale analyses will also be applied within this framework, to ensure the factors affecting in-basin variation are properly captured, and to provide scientists the capacity to apply their individual expertise to analyses. Furthermore, several individual studies will occur within the ecosystem indicators category, to analyze specific datasets for correlations with salmon survival.

Stoplight Example - W. Coast Vancouver Island - Trudel
Stoplight Example – W. Coast Vancouver Island – Trudel 2012.


Developing ecosystem indicators to inform geographic variation in marine survival rates

Investigators: Correigh Greene, Kathryn Sobocinski, (NOAA), Neala Kendall (WDFW)

Fisheries stocks are known for their high recruitment variability  and the resultant difficulty in forecasting stock size to support sustainable harvest management. In light of variation in marine survival, much interest has developed across the Pacific Coast in using ecosystem indicators to improve predictions of productivity and adult abundance. In Puget Sound, anadromous salmon stocks spawn in six sub-basins, each with its distinct geomorphology and oceanography, creating the potential for strong spatial structuring of productivity patterns. Coastal ecosystem indicators do not readily predict marine survival and productivity of stocks spawning within Puget Sound. Likewise, marine survival of stocks entering different basins of Puget Sound does not necessarily track survival rates from other basins. Therefore, the ability of ecosystem indicators to reduce uncertainty in forecasts of adult returns across Puget Sound appears to require local spatial variation.

This research activity produces and tests a spatially variable framework of ecosystem indicators that can be used by harvest managers to forecast adult salmon returns. Specifically, researchers are:

  • Developing a suite of indicators describing conditions within Puget Sound’s sub-basins as smolts enter seawater. This activity is being done in conjunction with Canadian indicators development for the Strait of Georgia.
  • Combining sub-basin indicators with ecosystem indicators of the Northwest region.
  • Testing these indicators for their ability to reduce uncertainty in predictions of marine survival.
  • Examining the potential for improved forecasts to reduce management constraints.
  • Developing tools that harvest managers in Puget Sound can use to forecast marine survival and adult returns.
  • Initiating plans for a workshop to show managers how these tools can be used.

Aggregation of candidate indicator datasets has been mostly completed. Qualitative network models were created to explore the relative influence of multiple ecosystem and anthropogenic drivers on salmon marine survival; results are pending. Further quantitative analysis of candidate indicators is underway.


Satellite-derived ecosystem indicators

Investigator: Brandon Sackmann (Integral Consulting)

This research activity leveraged historic and current satellite imagery to characterize long-term trends in near-surface water quality at a high spatial resolution in Puget Sound. Water quality indicators included sea surface temperature, water color, water clarity, algal biomass, and freshwater influence. These satellite-derived metrics provide sub-basin scale oceanographic data for Puget Sound and can be used to detect trends and potential impacts of regional change in the Salish Sea. The data products from this activity have been made available to other SSMSP projects and are being used to inform trend analyses and modeling efforts.


Analyzing Puget Sound JEMS zooplankton time series and other zooplankton data relative to physical controls and salmon survival

Calanus Copepod. Marine Zooplankton of Puget Sound ID card – Washington Sea Grant.

Investigator: Julie Keister (U. Washington)

Initial analysis of a zooplankton time series derived from the Joint Effort to Monitor the Strait (JEMS) sampling program revealed compelling correlations between variability in copepod species composition and coho salmon survival (2003-2009). The strength of correlation varied among salmon populations, but in many cases was strong enough to justify a need to better understand environmental controls on zooplankton. JEMS zooplankton data continue to be collected and analyzed, and annual updates to the coho survival time series are also underway – improving the power of ongoing correlational analyses. Further exploration of the JEMS time series for salmon prey indices that have a direct, mechanistic relationship to salmon survival is equally important; this investigation may lead to an index with a more rigorous and/or more intuitive relationship to salmon returns. This research activity builds on the initial JEMS analysis and extends its focus to also include a second, 5-year time series from Commencement Bay. The objectives of this analysis are to 1) identify rigorous metrics that will improve salmon return forecasting and 2) better understand bottom-up processes that drive ecosystem change in the Salish Sea.


Retrospective Analysis of Puget Sound Zooplankton Datasets

Investigators: Julie Keister (U. Washington) and Iris Kemp (LLTK)

Jellyfish. Marine Zooplankton of Puget Sound ID card – Washington Sea Grant.

This research activity explored a 20-year qualitative dataset of zooplankton presence near herring spawning areas across Puget Sound and detected broad-scale patterns of zooplankton community change. Results indicated both seasonal and interannual patterns in the zooplankton community, and potential sub-basin variation in zooplankton taxa presence; findings generally consistent with the hypotheses that Puget Sound sub-basins exhibit unique biological dynamics and that Puget Sound ecosystem dynamics have changed over time. Researchers identified an increase in presence of gelatinous organisms during January-April in South Puget Sound from the 1970s-1990s, and the data suggest shifts in presence and timing of several other zooplankton taxa. Read more here:


Linking Puget Sound primary production to stratification and atmospheric drivers

Investigator: Parker MacCready (U. Washington)

This research activity investigated the hypothesis that bottom-up pressures such as changes in phytoplankton and salmon prey availability have driven salmon marine survival declines. Time series of physical drivers (rivers, winds, sunshine), water conditions (stratification), and phytoplankton abundance and bloom timing were aggregated from USGS and Department of Ecology river flow datasets, Environment Canada data on Fraser River flow, weather data from SeaTac airport gauges, NCEP atmospheric data products, monthly Department of Ecology CTD casts, and daily data from ORCA profiling buoys. Time series aggregation and data analysis were limited, particularly given the lack of Puget Sound phytoplankton data. Clear correlations explaining interannual differences or trends in phytoplankton blooms were not found. However, results indicated that wind mixing may be a predictive variable for phytoplankton blooms, justifying further analysis in future. Read more here:


Linking Puget Sound plankton ecology to climate with a new, mid-complexity circulation model

Investigators: Neil Banas (U. Strathclyde) and Parker MacCready (U. Washington) 


Variation in Salish Sea river flow, in combination with climate-linked variation in cloud cover and winds, may drive variation in Puget Sound stratification. This may lead to changes in timing and magnitude of primary production, impacting the annual cycle of zooplankton taxa that are high-value prey for forage fish and early salmon life stages. This project couples a circulation model designed specifically for fjords like Puget Sound, a nutrient-phytoplankton-zooplankton model of primary production, and a trait-based model of copepod community structure and life cycle to understand the mechanistic linkages between climate and copepod populations in Puget Sound.

Example of Salish Sea model. Observed vs modeled results of chlorophyll in Puget Sound, reproducing scale of phytoplankton bloom – U. Washington.