LB 23-244 - CANCELLED - Meta-analysis of metabolic rates in three anadromous fishes reveals variation in scaling with mass, temperature and physical activity.

Background/Question/Methods:

Bioenergetics models of riverine fishes are a useful tool within ecosystem modeling for research and conservation. These models predict the energy available for fish to grow, migrate, reproduce, etc. using parameters calibrated from experimental data. However recent work in the field suggests more diversity in metabolic scaling relationships across species and across life stage than traditionally thought, and metabolic rate data are not always sufficiently available for parameter calibration across all the life stages of all species of concern. This can result in non-negligible errors in energy budgeting when ill-fitting parameters from one species/life stage are extrapolated to another. We performed a meta-analysis of three species of anadromous, protected fishes in North America with populations listed as Threatened or Endangered under the ESA: Chinook salmon (Oncorhynchus tshawytscha), steelhead (Oncorhynchus mykiss) and green sturgeon (Acipenser medirostris). To test whether bioenergetics models can be made more accurate by allowing greater flexibility in metabolic scaling relationships across life stage, we used two modeling methods to model energy consumption as a function of fish mass, swim speed, and temperature: a version of the traditional Wisconsin Model, and a generalized linear model (GLM) that incorporated interaction terms between fish mass and the other predictors.

Results/Conclusions:

In the two salmonid species, we found that allowing for interaction terms between mass and temperature and/or swim speed allowed our model to predict fish oxygen consumption across life stages more accurately and with more realistic residuals than the Wisconsin Model, particularly for some adults and small juveniles, for which the model predictions differed by up to ~70%. In green sturgeon, there was only marginal difference between the GLM and Wisconsin models in terms of quality of residual patterns, though predictions differed by as much as 50% at low temperatures, ~10ºC. Our findings reinforce the idea that scaling of metabolic rates with temperature, activity, and mass can differ across species and life stage. Moreover, we provide a modeling framework that incorporates interaction terms and greater flexibility to more accurately predict energy consumption. As environmental conditions (and thus, drivers of metabolic rates) shift with climate change for many species, it becomes increasingly important to ensure parameters for bioenergetic calculations are as accurate as possible to best inform conservation and management efforts.