1. Predicting the effects of rising temperature entails measuring both habitat thermal characteristics and the physiological variation of the species as it relates to this microhabitat variation; these two types of measurements can generate what is termed a “physiological landscape” for the species. Mapping the micro-scale physiological landscape across space and time, rather than relying on large-scale averages of temperature and means of thermal limits in a species, can allow more accurate estimates of an organism’s sensitivity to temperature change and support development of more refined models of the impacts of anthropogenic climate change that have higher predictive power.

2. We thus continually monitored the body (operative) temperature of the intertidal mussel Mytilisepta virgata in both sun-exposed and shaded microhabitats and determined the seasonal variations of cardiac performance of field-acclimatized and laboratory-acclimated mussels from different microhabitats for calculating the thermal sensitivity, as indicated by the difference between the maximum ambient temperature and an individual’s upper thermal limit (thermal safety margin, TSM), in each microhabitat and each month.

3. The mussels experienced divergent thermal stress, in average temperature, acute and chronic thermal stress, and thermal predictivity among different microhabitats, and presented high spatial-temporal variations of cardiac function as results of seasonal acclimatization and inter-individual variations. Results of TSMs indicated that the thermal sensitivities of the mussels to high temperature were season- and microhabitat-specific, and the mussels in the shaded microhabitats were predicted to survive the hottest summer temperatures; however, some individuals in the sun-exposed microhabitats experienced temperatures above their sublethal temperature.

4. With the large, high-resolution dataset of thermal environmental characteristics and the cardiac performances with high variations, we were able to integrate the effects of synchronized changes in microenvironmental temperatures and cardiac thermal responses and thereby characterize the physiological landscape of thermal sensitivity. The complex physiological landscape that exists in the intertidal zone must be taken into account when predicting effects of changes in environmental temperature, such as those occurring with global climate change.

The operative temperature of the mussels was continuously recorded using biomimetic thermal loggers, Robo-mussels. For preparing a Robo-mussel, a data logger iButton (DS1922L, Maxim Integrated, CA, USA) was inserted into a shell of a mussel (shell length, 4-5 cm) filled with 3M Scotchcast 2130 Flame Retardant Compound as described in the previous study. A total of 12 Robo-mussels were deployed at six microhabitats at the Dongshan Swire Marine Station (D-SMART), Fujian Province, China (23.65°N, 117.49°E). In each microhabitat, two Robo-mussels were deployed between 1.80 and 2.50 meters above chart datum (CD, the level from which depths displayed on a nautical chart are measured) where the black mussel normally occurs. Among the six microhabitats, three were south-facing (sun-exposed) and three were north-facing (shaded) microhabitats. After May 2019, due to the low density of the mussel in the shaded #3 microhabitat, another shaded microhabitat (shaded #4) was added as a substitute. Temperature was recorded at an interval of 1 hour and an accuracy of 0.5°C from May 2018 to September 2019.

A non-invasive method was used to measure individual heart rates. An infrared sensor (IR-EX, Newshift, Portugal) for heart rate measurement was directly attached above the mussel's peripheral heart cavity using Blu-Tack (Bostik, Wauwatosa, Wisconsin). The mussel was heated in air using a water bath (TXF 200, Grant, UK) from 25°C at a rate of 6°C h-1 until a temperature was reached where the heart rate fells to zero (Flat line temperature, FLT) as described in our previous study. Body temperature was measured simultaneously during heating using thermocouple thermometers inserted inside the shell (Fluke 54II, Fluke, WA). The infrared heartbeat signal was amplified (AMP03, Newshift, Leiria, Portugal), filtered, smoothed, and recorded with a PowerLab unit (16/30, ADInstruments, March-Hugstetten, Germany). Data were viewed and analyzed using LabChart (v. 7.2) (ADInstruments, Germany). Arrhenius break temperature (ABT) values were calculated by using linear regression in Origin ver. 9 (OriginLab Corporation, MA, USA).