2012 Wild Coho Forecasts for Puget Sound, Washington Coast, and Lower Columbia

Run size forecasts for wild coho stocks are an important part of the pre-season planning process for Washington State salmon fisheries. Accurate forecasts at the level of management units are needed to ensure adequate spawning escapements, realize harvest benefits, and achieve harvest allocation goals.

Wild coho run sizes (adult ocean recruits) have been predicted using various approaches across Washington’s coho producing systems. Methods that rely on the relationship between adult escapement and resulting run sizes are problematic due to inaccurate escapement estimates and difficulty allocating fishery catches by stock. In addition, escapement-based coho forecasts often have no predictive value because watersheds become fully seeded at low spawner abundances (Bradford et al. 2000). Furthermore, different variables in the freshwater (Lawson et al. 2004; Sharma and Hilborn 2001) and marine environments (Logerwell et al. 2003; Nickelson 1986; Ryding and Skalski 1999) influence coho survival and recruitment to the next life stage. Therefore, the accuracy of coho run size forecasts should be improved by partitioning recruitment into freshwater production and marine survival. In this forecast, wild coho run sizes (adult ocean recruits) are the product of smolt production and marine survival and are expressed in a matrix that combines these two components. This approach is similar to that used to predict hatchery returns where the starting population (number of smolts released) is known.

Freshwater production, or smolt abundance, is measured as the number of coho smolts leaving freshwater at the conclusion of the freshwater life stage. The Washington Department of Fish and Wildlife (WDFW) and tribal natural resource departments have made substantial investments in monitoring smolt populations in order to assess escapement goals and improve run size forecasts. Long-term studies on wild coho populations have been used to identify environmental variables contributing to freshwater production (e.g., low summer flows, pink salmon escapement, watershed gradient). For stocks where smolt abundance is not measured, smolt production is estimated by using the identified correlated to extrapolate information from neighboring or comparable watersheds.

Marine survival is survival from saltwater entry through the ocean rearing phase to the point that harvest begins. Marine survival for a given stock is measured by summing coho harvest and escapement and dividing by smolt production. Marine survival rates for wild coho stocks have been measured at four stations in Puget Sound and at one station in the Grays Harbor system. Harvest of wild coho produced by these watersheds is measured by releasing a known number of coded-wire tagged wild coho smolts and compiling their recoveries in coastwide fisheries. Coastwide recoveries are compiled from the Regional Mark Processing Center database (www.rpmc.org). Tags in returning spawners are enumerated at upstream trapping structures. Results from these monitoring stations describe patterns in survival among years and watersheds. These patterns are used to predict marine survival of the wild coho cohort that is currently recruiting into the fisheries. For stocks where marine survival is not measured, adult ocean recruits are predicted by extrapolating information from neighboring or comparable watersheds.

The Wild Salmon Production Evaluation (WSPE) Unit within the WDFW Fish Program Science Division has developed forecasts of wild coho run size for the last seventeen years. Beginning in 1996, a wild coho forecast was developed for all primary and most secondary management units in Puget Sound and the Washington coast (Seiler 1996). A forecast methodology for Lower Columbia wild coho was added in 2000 (Seiler 2000) and continued to evolve in response to listing of Lower Columbia coho under the Endangered Species Act in 2005 (Volkhardt et al. 2007).

Table 1 summarizes the 2012 run-size forecasts for wild coho for Puget Sound, Washington Coast, and Lower Columbia River systems. Forecasts of three-year old ocean recruits were adjusted to January age-3 recruits in order to provide appropriate inputs for coho management models (expansion factor = 1.23, expansion provides for natural mortality). December age-2 recruits, which have been included in this table in previous years, are not provided as they are no longer used by fisheries managers. The following sections describe the approach used to derive smolt production and marine survival estimates.