Winter in the Pacific Northwest is characterized by frequent rainfall and southwesterly winds. Southwest winds push water onshore and cause downwelling (the opposite of upwelling). Downwelling in turn brings warm, nutrient–depleted, surface water onshore from offshore sources and results in very low levels of primary production. The most critical time of the seasonal plankton–production cycle is when the ocean transitions from a winter downwelling state to a summer upwelling state. This time is known as the spring transition.
The spring transition marks the beginning of the upwelling season and can occur at any time between March and June. Generally, the earlier in the year that upwelling is initiated, the greater ecosystem productivity will be in that year. In some years the transition is sharp, and the actual day of transition can be identified easily, but in many years transition timing is more obscure. It is not uncommon for northerly winds (favorable to upwelling) to blow for a few days, only to be followed by southwesterly winds and storms. Intense, late–season storms can erase any upwelling signature that may have been initiated, thus re–setting the "seasonal clock" to a winter state. This is what occurred during summer 2005.
Anomalies in the date of the physical spring transition from 1969 to present. Anomaly is based on an average date of 13 April using the minimum cumulative upwelling index (CUI) value.
The date of spring transition can be indexed in several ways. Here, we use the date of the minimum value of the Cumulative Upwelling Index (CUI). Further details can be found in Bakun (1973) and Bograd et al. (2009). The average date of upwelling is 13 April (Day 103), but can range from early March to early June. Note from Figure PST-01 the following points:
- Most spring transition dates during the pre–1977 cool–phase PDO were earlier than average.
- Spring transition dates from the 1980s and 1990s did not reflect changes in either the PDO (Figure PDO-01) or the Multivariate ENSO index (Figure ONI-01).
- The period of early transition dates from 1985 to 1990 correlates well with the high salmon survival in the late 1980s (see Figure PDO-02).
We have developed a new measure of the spring transition based on measurements of temperature taken during our biweekly sampling cruises off Newport, Oregon. We define the spring transition as the date on which deep water colder than 8°C was observed at the mid shelf (station NH 05). This indicates the presence of cold, nutrient–rich water that will upwell at the coast with the onset of strong northerly winds, signaling the potential for high plankton production rates.
Figure PST-02 (left) shows that adult spring and fall Chinook returns at Bonneville and coho salmon smolt-to-adult survival from 1996 to present are not well correlated with the day of spring transition using the Cumulative Upwelling (Bakun) method, though studies using an earlier and longer time series have found some correlation (Logerwell et al. 2003). An analysis using smolt–to–adult return rates of Snake River spring/summer Chinook salmon (from Scheuerell and Williams 2005) did not reveal any significant correlations with the transition date.
However, the spring transition date using our new hydrographic method does show a weak, but significant relationship with adult returns of adult spring and fall Chinook returns at Bonneville, and coho salmon smolt-to-adult-survival (Figure PST-02, right). Survival is higher in years with an earlier transition date.
Figure PST-02. Plot showing the relationship between spring Chinook salmon adult returns at Bonneville dam (lag 2 years, top), fall Chinook salmon adult returns at Bonneville dam (lag 2 years, middle), and coho salmon smolt-to-adult survival (SAR, lag 1 year, bottom) versus the date of spring transition using the Bakun upwelling method (left) and the hydrographic method (right). Numbers indicate the warm (red) and cold (blue) years. Years in black were outliers and were excluded from the regression.
*outliers were excluded using Cook's Distance
PST-02 shows that hatchery adult coho salmon returns from 1996 to present are not well correlated with the day of spring transition, though studies using an earlier and longer time series have found some correlation (Logerwell et al. 2003). An analysis using smolt–to–adult return rates of Snake River spring/summer Chinook salmon (from Scheuerell and Williams 2005) did not reveal any significant correlations with the transition date, but when using spring Chinook salmon adult counts at Bonneville Dam (1996 – present from the
DART) there is a significant linear relationship (PST-02).
Other measures of the spring transition include ones from:
- Dr. Mike Kosro, College of Earth, Ocean and Atmospheric Sciences (CEOAS), Oregon State University, who operates an array of coastal radars that are designed to track the speed and direction of currents at the sea surface. He produces
daily charts showing ocean surface current vectors, and from those one can clearly see when surface waters are moving south (due to upwelling) or north (due to downwelling). By scanning progressive images, the date of transition can be visualized.
- Dr. Steve Pierce and Dr. Jack Barth, CEOAS, Oregon State University, use local wind data from Newport, Oregon and produce
annual plots of the start and end to the upwelling season based on the change in alongshore windstress.
- Logerwell et al. (2003) indexed the spring transition date based on the first day when the value of the 10–day running average for upwelling was positive and the value of the 10–day running average for sea level was negative. This index is no longer regularly updated and made available on-line.