Estimating thermal performance curves from repeated field observations

By:  and 



Estimating thermal performance of organisms is critical for understanding population distributions and dynamics and predicting responses to climate change. Typically, performance curves are estimated using laboratory studies to isolate temperature effects, but other abiotic and biotic factors influence temperature-performance relationships in nature reducing these models' predictive ability. We present a model for estimating thermal performance curves from repeated field observations that includes environmental and individual variation. We fit the model in a Bayesian framework using MCMC sampling, which allowed for estimation of unobserved latent growth while propagating uncertainty. Fitting the model to simulated data varying in sampling design and parameter values demonstrated that the parameter estimates were accurate, precise, and unbiased. Fitting the model to individual growth data from wild trout revealed high out-of-sample predictive ability relative to laboratory-derived models, which produced more biased predictions for field performance. The field-based estimates of thermal maxima were lower than those based on laboratory studies. Under warming temperature scenarios, field-derived performance models predicted stronger declines in body size than laboratory-derived models, suggesting that laboratory-based models may underestimate climate change effects. The presented model estimates true, realized field performance, avoiding assumptions required for applying laboratory-based models to field performance, which should improve estimates of performance under climate change and advance thermal ecology.

Publication type Article
Publication Subtype Journal Article
Title Estimating thermal performance curves from repeated field observations
Series title Ecology
DOI 10.1002/ecy.1801
Volume 98
Issue 5
Year Published 2017
Language English
Publisher Ecological Society of America
Contributing office(s) Leetown Science Center
Description 11 p.
First page 1377
Last page 1387
Google Analytic Metrics Metrics page
Additional publication details