The endangered status of different, globally distributed, salmon populations makes it increasingly important to understand the factors that influence their successful reproduction. Salmon are anadromous species and spend different parts of their life cycle in both the oceanic and the land [river] system. So environmental factors affecting their development may range from genetic to variations in river-runoff and associated acid-rain levels to ocean currents and climatic changes (e.g. sea temperatures). This pilot project examines whether there is any detectable transfer of information/climate signal from land-river systems to air temperatures and hence whether basic autocorrelation methods could produce climatic indices relevant to anadromous species?

LAND - CANADA
Many northern river systems have a strong spring runoff potentially associated with climatic variations. If marine signatures can be detected in the atmosphere during the autumn seasonal transition, would the spring season be associated with any feedback [mechanism] between river/land water systems and the atmosphere? The MONACLE technique was used to generate a regional, 4-month autocorrelation series over the spring months. The resulting indices exhibited visual comparisons with times series of salmon catch at approx. 95% significance levels but with moderate to low levels of explained variance. A bootstrap re-sampling analysis indicated these relationships were more than accidental.

Salmon catch (red) at St. John's [NB, Canada] and a local spring climate indicator (black).

Salmon catch (green) for British Columbia [Canada] and a local spring climate indicator (black).

OCEANS - ALASKA
Other authors have previously described the relationship between Alaska salmon catch and oceanic processes in the Alaskan gyre region of the N. Pacific. This known oceanic link may provide an understanding of why the Alaska salmon catch data [blue] shows a better comparison with an autumn (marine-related) climate indicator [black] rather than a possible land/river-related spring indicator. So could obtaining matches for the environment-to-an-index at different seasons, with MONACLE, indicate changing land-ocean influences for these species?

Alaska salmon catch (blue) and an autumn climate indicator (black).

Since catch shows a relation to regional climate indicators, an exploration of basin-scale processes was conducted using the longest historic catch as a quasi-index.

HISTORIC MAP [1880-1970]
This figure shows locations where long term variations in a spring climate indicator are correlated to historic Salmon catches in the Saint John [NB] area. [Red positive match, blue-negative match, grey areas data available but no match]. The coastal regions with similar [red] or inverse [blue] trends have, historically, been associated with major salmon fishing efforts. [Colour scales are set at the 99% sig. level].

Salmon-climate index relationship for 91 years (1880-1970).

RECENT YEARS: 1970s on
The last two decades has seen a dramatic decline in salmon stock on both sides of the North Atlantic. Since several natural populations are no longer commercially fished any climate relationship has been investigated using a hypothetical, exponentially declining population from the 1970s on.

Relationship for Salmon to exponential time trend “index” for 1970 on.

This figure shows regions in southern BC and Scandinavia (red) do have a positive comparison but there is no evidence of the prior match with the inland, Great Lakes region. There is an inverse [blue] relationship around the northern North Atlantic [Labrador, Iceland, and UK]. No causal relationships are suggested but this pattern could indicate links to recent North Atlantic climatic processes and have relevance for climate scenario studies.

SUMMARY
Statistical relationships alone provide minimal insight into environmental processes. Whether the above spring index approach remains robust is still to be determined. Both rainfall and ocean temperature data should be incorporated into this study. The questions of whether spring atmospheric data contains information/ feedback from the climate system in a similar manner to autumn data, or whether the role of ocean vs. land influences have changed, continues to be explored.

ACKNOWLEDGEMENTS:
Data has been taken from numerous reports and publications plus the databases of NOAA, FAO, NAFO, NMFS and DFO. Special thanks to Dr. Larry Marshall, and Peter Amiro with Diadromous Fish Division at BIO for data and numerous helpful discussions.