Springs and spring-influenced streams in aridland riverscapes are important harbors of endemic faunas, yet they are threatened by land use change, groundwater extraction, and climate disruption. Despite these anthropogenic stressors, surprisingly few spatiotemporally complete temperature datasets exist for these habitats, limiting our understanding of baseline thermal regimes and our ability to forecast changes. To address this gap, we monitored temperature variation at 15-minute intervals for 24 months at 147 monitoring stations distributed along a gradient of spring influence and urbanization in central Texas, USA. This monitoring network generated > 4 million water temperature measurements. The sensitivity of water temperature to diel and seasonal air temperature fluctuations varied substantially across monitoring stations and increased at lower-magnitude springs and with increasing downstream distance from springs. Statistical models fit with spatial (e.g., distance from spring) and temporal (e.g., air temperature) covariates predicted daily maximum water temperature with high accuracy (r2=93). Next, we used laboratory assays to measure upper thermal tolerances at acute and chronic timescales and across acclimation temperatures for six putative spring-associated fishes, including one state-listed and one federally-listed species. These assays revealed substantial interspecific variation in acute tolerance, chronic tolerance, and acclimation capacity. Integrating field-based thermal exposure with lab-based thermal sensitivity suggests that springs provide essential thermal refugia for the most thermally sensitive species, whereas other species exhibit tolerance to withstand higher temperatures beyond the realm of spring influence. Overall, our findings reveal substantial spatial variation in thermal exposure and taxonomic variation in thermal sensitivity that, together, can inform spatially-explicit and species-specific conservation actions.