Improved thermal preferences and a stressor index derived from modeled stream temperatures and regional taxonomic standards for freshwater macroinvertebrates of the Pacific Northwest, USA
Benthic macroinvertebrate taxa vary in their sensitivities to water quality and habitat conditions, contributing to their extensive use as ecological indicators. As climate change and landscape alteration increasingly impact stream temperatures, interest is growing in expanding our knowledge of how macroinvertebrates are affected by current and future thermal conditions. Using samples from 3501 sites, we evaluated relationships between macroinvertebrate taxa and modeled stream temperatures across Oregon and Washington, in the U.S. Pacific Northwest. We used Maximum Weekly Maximum Temperature (MWMT) values from the NorWeST temperature dataset, which is the same metric used for numeric water temperature standards in Oregon and Washington. MWMT captures peak thermal stress, when cold-water adapted aquatic biota are closest to their upper physiological limits. For each macroinvertebrate taxon, we characterized relationships between MWMT and their distributions with three measures: 1) central thermal tendency, based on weighted average (WA) optima calculations and relative abundance data; 2) lower and upper thermal limits, based on the 10th and 90th percentiles of taxon occurrence, using presence data; and 3) thermal sensitivity curve shape, based on Generalized Additive Model (GAM) plots. We assigned 521 taxa, from species to phyla, to seven thermal preference categories, ranging from cold and warm stenotherms (narrow range) to eurythermal (wide range). Thermal sensitivity and variability within each taxonomic group were identified for establishing taxonomic targets for regional monitoring programs. We also developed the Macroinvertebrate Thermal Tolerance Index (MTTI) to represent the assemblage-level response to available thermal habitats, using WA optima and relative abundances for 324 taxa. The MTTI model had a strong relationship with modeled temperatures (R2 = 0.68) and a root-mean-square-error of 2.5 °C. Our work builds on previous regional and national efforts to identify thermal indicator taxa by using modeled stream network temperatures and a thermal metric that corresponds directly to regional water temperature standards. Both the taxa thermal preferences and the MTTI can be used to help identify causes of biological impairment, prioritize restoration and protection actions, and monitor assemblage-wide changes in thermal tolerance over time.