Results and Discussion

Results

Weak Correlation

Our data indicate that increasing temperatures are negatively influencing salmon Recruits per Spawner, yet positively influencing salmon Returns (Table 2). However, very minor influences were observed: a weak negative correlation was found between the number of Recruits per Spawner (R.S) and Mean Annual Temperature (MAT), with MAT explaining less than 1.1% of the variation in R.S (Table 2, Figure 12). Similarly, a weak negative correlation was observed between the Recruits per Spawner and Mean Annual Precipitation (MAP), with MAP explaining 0.1% of the variation in R.S (Table 2, Figure 13). A weak positive correlation was observed between the number of Returns and MAT, with MAT explaining 1.6% of the variation in salmon returns, and a weak positive correlation was observed between the number of returns and MAP, with MAP explaining 0.8% of the variation in salmon returns (Table 2, Figures 9 & 10).

The slopes reiterate that very minor influences were observed, such as a decrease of ~1 Recruit/Spawner and an increase of ~1 Return for every degree increase in temperature (Table 2).

Table 2. Output from Correlation and Regression analysis performed in R software comparing salmon Returns and the number of Recruits per Spawner (R.S) as a function of Mean Annual Temperature (MAT) and Mean Annual Precipitation (MAP), as well as R.S, Returns, MAT, and MAP as a function of Year. The slope values were back-transformed into original units as this analysis involved taking the log of the Recruits per Spawner (R.S) and Returns.

At the species level, the trend of a slight decrease in Recruits per Spawner in response to rising Mean Annual Temperature (MAT) and Mean Annual Precipitation (MAP) was similar for all 4 species, while trends in salmon Returns were highly variable among these species (Figures 12 & 13). Increasing MAT and MAP appeared to have a slight negative influence on Pink salmon Returns, though Sockeye and Chum salmon demonstrate the opposite trend (Figures 12 & 13).

Figure 12. Salmon productivity metrics (Returns and Recruits per Spawner) in response to Mean Annual Temperature (MAT), with linear regression lines for each species. Legend applies to both plots. Note that 'Pink even' and 'Pink odd' refer to the same species of Pink salmon separated by the year they returned to spawn.

Figure 13. Salmon productivity metrics (Recruits per Spawner and Returns) in response to Mean Annual Precipitation, with linear regression lines for each species. Legend applies to both plots. Note that 'Pink even' and 'Pink odd' refer to the same species of Pink salmon separated by the year they returned to spawn.

These observations were further explored using a linear model with Area as a block treatment (Tables 3, 4, 5 & 6). Mean Annual Temperature (MAT) was found to have a significant effect on the Recruits per Spawner [F(1,2653)=20.07, p=<0.0001], while Mean Annual Precipitation did not [F(1,2653)=3.71, p=0.0542] at an alpha level of 0.05. Both MAT and MAP had a significant effect on the number of Returns [F(1,3357)=59.34, p=<0.0001], [F(1,3357)=29.41, p=<0.0001] respectively (Tables 5 & 6).

For the number of Recruits per Spawner (R.S), 75.1% of the variance observed was explained by the Mean Annual Temperature (MAT), while only ~8% was explained by the Area block 'treatment' (Table 3). However 75.3% of the variance in Returns observed was explained by the Area, and only 24.1% of the variance was explained by the MAT (Table 5). Mean Annual Precipitation was largely determined by the Area as 59.3% of the variance observed in R.S and 81.7% of the variance in Returns was due to the Area block 'treatment' (Tables 4 & 6).

Table 3. Output from the linear model comparing the log(Recruits per Spawner) to the Mean Annual Temperature. Includes an additional column (%SS), which demonstrates what proportion of the variance in Recruits per Spawner is explained by the Mean Annual Temperature.

Table 4. Output from the linear model comparing the log(Recruits per Spawner) to the Mean Annual Precipitation. Includes an additional column (%SS), which demonstrates what proportion of the variance in Recruits per Spawner is explained by the Mean Annual Precipitation.

Table 5. Output from the linear model comparing the log(Returns) to the Mean Annual Temperature. Includes an additional column (%SS), which demonstrates what proportion of the variance in Returns is explained by the Mean Annual Temperature.

Table 6. Output from the linear model comparing the log(Returns) to the Mean Annual Precipitation. Includes an additional column (%SS), which demonstrates what proportion of the variance in Returns is explained by the Mean Annual Precipitation.

A Changing Climate

Climate data retrieved for the years 1950-2013 demonstrate that the average temperature has clearly risen by an average of 0.019 degrees C per year (Figure 14). While precipitation has decreased by an average of 0.136 of a millimeter (Figure 15).

Figure 14. Temperature response variables (Mean Annual Minimum, Mean Annual Average, and Mean Annual Maximum Temperature) recorded between 1950-2013.

Figure 15. Precipitation Response Variable (Mean Annual Precipitation) recorded between 1950-2013.

Discussion

Temperature appeared to have a negative influence on annual recruits per spawner, suggesting a negative effect of higher temperatures on successful spawning. This finding was corroborated by Whitney (2012), who also found that water temperatures can influence salmon gametogenesis. Additionally, increased precipitation appeared to have a slightly positive influence on the annual rate of Returns. While the influence of climate on recruits per spawner was very similar for all species, the returns varied widely across the three species.

The methodology and limited results of this study can be used in the future to reevaluate trends in salmon productivity and climate. Considering salmon are incredibly sensitive to changes in their environment, it is imperative to continually monitor their spawning and return rates along the coast as climate data from this time period clearly demonstrated that temperatures are rising, while precipitation may be decreasing (Whitney, 2012). These findings are corroborated by Patterson et al. (2007), who found that the mean annual daily temperature in the Fraser River, BCM had risen 1.3°C since the 1950s upon completion of their study in 2007. Additionally, while only minor species-specific trends were observed in this study, further analysis is required as prior research has demonstrated certain species of salmon may be more resistant to climatic changes than others (Brett, 1952).

There were several limitations in our data that may have impacted the accuracy of our analysis. Of particular note is the limited data on locations that required manual input of coordinates through Google Maps, which did not reflect actual sampling locations and was instead used primarily to link our climate data to the raw information collected by Ogden et al. (2015). This study was also the first known attempt to develop a productivity rating for Pacific salmon and is therefore characterized by large uncertainties in data quality (Ogden et al. 2015). Additionally, as with any data relying on surveys, survey effort was noted by Ogden et al. (2015) as a source of potential error in the raw dataset.

For future studies, an emphasis on consistent data collection and clear directives for sampling procedures and locations would be an asset. Data on water temperatures could also be incorporated over the same time period, as air temperatures can vary more widely than water temperatures (Webb & Nobilis, 1997). As well, other metrics of salmon productivity, such as size and weight, and spawning/migration period (Patterson et al., 2007) may offer additional insight as to whether Pacific salmon are affected by the changing climate.

Conclusion

Increasing temperatures resulting from climate changes have been shown to impact the life cycles of various fish species (Whitney, 2012). The primary findings of this study suggest the climatic changes that occurred during the study period had very little influence on the three salmon species analyzed; Increased temperature appeared to slightly negatively influence salmon recruitment, while slightly positively influencing salmon returns. The minor trends observed could indicate that climatic changes are affecting certain aspects of salmon lifecycles, particularly when coupled with numerous other factors, such as pollution, disease, and angling pressure that were not accounted for in this study. Nevertheless, this study is valuable in its contribution to the understanding of environmental influences on Pacific salmon and should encourage further research on potential threats to these critical species.