Forecasts and Planning Tools

Seasonal to Interannual Forecasts

Forecasting Oregon Coho Marine Survival

May 2008 archive copy

People Involved

• Elizabeth Logerwell (contact person), NOAA Fisheries/Alaska Fishery Science Center
• Nathan Mantua, University of Washington, Climate Impacts Group
• Peter Lawson, NOAA Fisheries/Newport Lab
• Robert Francis, University of Washington, Climate Impacts Group and School of Aquatic and Fishery Sciences
• Vera Agostini, University of Washington, Climate Impacts Group and School of Aquatic and Fishery Sciences

Background

Marine survival rates for Oregon coho salmon are influenced by several sequential environmental processes that affect coastal ocean food webs:

1. Winter climate prior to smolt migration from rivers to the ocean;
2. Date of occurrence of the “spring transition”, the period when alongshore winds (off the Pacific Northwest (PNW) coast) shift from being mostly northward to mostly southward;
3. Total coastal upwelling during the spring;
4. Ocean conditions during the maturing cohos’ only winter at sea.

By monitoring conditions during each cohort’s lifecycle, we can provide an experimental forecast of Oregon coho marine survival rates.

Recent and Forecasted Conditions

Current Forecast

Issue date: May 8, 2008

FORECAST RETURN RATE FOR ADULTS RETURNING FALL 2008:

Ocean conditions, as measured by our simple model, were better than average for OPI coho smolts that entered the ocean in 2007. The spring transition was near average and spring sea level was low (low sea level is indicative of good upwelling and strong north-south transport). January through March SST in 2008 was among the coldest on record and strongly favorable.

Based on these favorable conditions this model projects return rate of 5% (+/- 1%) for fall 2008. However, jack (2-year-old male) returns in the fall of 2007 to the Columbia River were at low levels typical of the 1990s, resulting in a projected return rate of 0.9%. It was also notable that upwelling in June, July and August 2007 were weak, and coastal ocean temperatures failed to cool as they typically do. These summer conditions are not considered in our simple model, but may help to explain the discrepancies between our model forecast and the jack-based forecast that calls for much lower return rates.

Past Forecast Performance

Forecast Methodology

To better understand and predict Oregon coho marine survival, we developed a conceptual model (Figure 2) of key environmental processes that influence coastal ocean food webs and ultimately marine survival rates for Oregon coho salmon. The key processes are sequential:

1. winter climate prior to smolt migration from rivers to the ocean;
2. date of occurrence of the spring transition, when alongshore winds (off the PNW coast) shift from being mostly northward to mostly southward;
3. total coastal upwelling during the spring;
4. ocean conditions during the maturing cohos’ only winter at sea.

We then parameterized a general additive model (GAM) with Oregon Production Index (OPI) coho smolt-to-adult survival estimates from 1970-2001 and the environmental processes listed above. For the model training period (smolt year data from 1969-2000), the GAM explained 75% of the variance in observed OPI smolt-to-adult survival rates.

Selected References

Logerwell, E. A., N. J. Mantua, P. Lawson, R. C. Francis, and V. Agostini. 2003. Tracking environmental processes in the coastal zone for understanding and predicting Oregon coho (Oncorhynchus kisutch) marine survival. Fisheries Oceanography 12(3):1-15.