Assessing Patterns of Resilience to Climate Change for Early-Run Chinook Across Latitudes

Chinook salmon display high life-history diversity and capacity for behavioral acclimation and adaptation to local conditions, allowing them to occupy a number of freshwater habitats near their thermal limits. The pattern and extent of vulnerability of Chinook to future climate change is not well understood; especially accounting for the joint effects of heterogeneous climate impacts and population dynamics that can potentially buffer the fish from temporally and spatially variable mortality.  For a series of focal populations from major river basins spanning the NE Pacific range of Chinook salmon, we used a life-cycle modeling approach to ask: (1) what are projected population trends under future warming? And (2) what specific life stages are most sensitive to increases in freshwater temperatures? We constructed a life-cycle dynamics model using functional relationships between temperature and stage-specific survival during egg, juvenile, adult migration, and prespawner stages. We modeled 10 populations from interior tributaries to the Sacramento, upper Columbia, lower Thompson/Fraser, and Yukon Rivers where adult escapement, timing, harvest, age structure, and juvenile abundance data were available. For each population, we parameterized a multistage Beverton-Holt model, with dynamics influenced by spatially-explicit temperature impacts on four freshwater stages, and under (+/-) PDO periods of assumed ocean survival. We ran a series of Monte Carlo simulations with uniform increments of warming above the baseline temperature distribution experienced by cohorts. Population responses to modeled warming are varied within major river basins as well as across latitude.  For example, projected population productivities (log_e(R/S)) fall below replacement rates after the smallest temperature increases in the Chena (Yukon Basin), Coldwater (Fraser Basin) and upper Wenatchee (Columbia Basin) Rivers, where temperature unrelated factors may already depress productivity rates.  In the Sacramento, Nicola and Wenatchee watersheds, low elevation mainstem populations display higher abundance and productivity than nearby spring-run populations at higher spawning elevations, despite facing greater near-term thermal stress and greater projected rates of decline. The 3 populations in the Sacramento River Basin range from relatively low vulnerability to increased temperature (Mill Creek) to high vulnerability (Butte Creek). The stages most sensitive to future temperature increases also are heterogeneous throughout the river basins we modeled.  For example, changes in population productivity are most sensitive to temperature-induced mortality at the migration stage for the Yukon Basin populations, the prespawner stage for all 3 populations from the Sacramento River; and are equally sensitive to migration/prespawner stage temperatures for the lower Wenatchee River (Columbia Basin) summer-run population.