The ecological importance of the accuracy of environmental temperature measurements.

Abstract

The implications of logger accuracy and precision are rarely considered prior to their application in many ecological studies. We assessed the accuracy and precision of three temperature data loggers widely used in ecological studies (Hobo®, iButton® and TinyTag®). Accuracy was highest in TinyTags (95% of readings were within 0.23°C of the true temperature) and lowest in HOBOs and iButtons (95% of were readings within 0.43°C and 0.49°C of the true temperature, respectively). The precision (standard deviation of the repeat measurements) was greatest in TinyTags (0.04°C), followed by iButtons (0.17°C) and then HOBOs (0.22°C). As a case study, we then considered how modelled estimates of sea turtle hatchling sex ratios (derived from temperature), could vary as a function of logger accuracy. For example, at 29°C when the mean sex ratio derived was 0.47 female, the sex ratio estimate from a single logger could vary between 0.40 and 0.50 for TinyTags and 0.29 and 0.56 for both HOBOs and iButtons. Our results suggest that these temperature loggers can provide reliable descriptions of sand temperature if they are not over-interpreted. Logger accuracy must be considered in future ecological studies in which temperature thresholds are important.