Microclimate Temperature Sensor Performance Is Good but Boosted by a Simple Calibration Against a Research‐Grade Standard
- 1Institute of Botany and Landscape Ecology University of Greifswald Greifswald Germany
- 2Remote Sensing and Geoinformatics GFZ German Research Centre for Geosciences Potsdam Germany
Abstract
Small, self‐sufficient and affordable temperature loggers are widely used in ecological and biogeographical studies, but their accuracy, precision and possible lags relative to research grade sensors are incompletely quantified; simple, field applicable calibration procedures are needed to improve data quality and comparability.
We evaluated four commonly used low‐cost loggers (HOBO U23 001A, iButton DS1925L F5#, TOMST TMS 4, TOMST Thermologger) against a Campbell Scientific EE181 L reference in a climate chamber (PERCIVAL LT 36VLX). Three replicates per logger type recorded at 5‐min intervals across six 24‐h cooling–warming cycles. The chamber was set to cycle between −10°C and +35°C; however, the sensors recorded an actual achieved range of −11.5°C to +31.5°C. Data were binned in 5°C steps to assess accuracy and precision across the gradient; we tested for time lags and derived linear, sensor specific self‐calibration formulas of the form T adj = T sensor − ( a + b × T sensor ).
All loggers remained within ±1°C of the reference across the tested range, with larger errors at cold temperatures and improved performance at warmer temperatures. The EE181 L showed within‐sensor precision < ±0.1°C at ~23°C; low‐cost loggers achieved within‐sensor precision of ±0.10°C–0.15°C. All low‐cost devices exhibited short time lags of up to a few minutes in tracking temperature changes; a 5‐min timestamp shift compensated for the lag of HOBO sensors. Applying linear self‐calibration substantially reduced RMSE for all devices (e.g., iButton RMSE 0.46°C → 0.08°C; HOBO 0.22°C → 0.16°C) and improved accuracy to at least ±0.25°C across temperatures.
Common low‐cost loggers provide cost effective microclimate data with high precision and adequate accuracy for many field studies but show systematic, temperature dependent offsets and short temporal lags that warrant correction. We recommend sensor and deployment specific calibrations.
Versions
Related
- Geometry-informed temperature band screening for crop microclimate management in a commercial tomato greenhouse
- How body size and microclimate determine when solar absorptivity matters for insect body temperature
- Diagnostic Performance of Immunoglobulin E Serology, Prick Test and Patch Test With Food Allergens in Dogs With Atopic Dermatitis and Adverse Food Reactions Using the Double‐Blind Placebo‐Controlled Food Challenge as Gold Standard
- Assessing measurement uncertainty at laboratory network scale and its impact on diagnostic performance in the absence of a gold standard: application to ELISA tests for Coxiella burnetii in ruminants
- Microclimate Comfort of Elevated Bridge Underpass Spaces Based on Sound and Thermal Environments: A Case Study of Fuzhou and Xiamen, Fujian Province
BibTeX
@article{harnisch2026microclimate,
title = {{Microclimate Temperature Sensor Performance Is Good but Boosted by a Simple Calibration Against a Research‐Grade Standard}},
author = {Daniel Harnisch and Marko Smiljanić and Steffen Schulz and Joraine Schmoldt and Martin Wilmking},
journal = {Ecology and Evolution},
year = {2026},
month = nov,
doi = {10.1002/ece3.74733},
url = {https://onlinelibrary.wiley.com/doi/10.1002/ece3.74733},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Microclimate Temperature Sensor Performance Is Good but Boosted by a Simple Calibration Against a Research‐Grade Standard
AU - Daniel Harnisch
AU - Marko Smiljanić
AU - Steffen Schulz
AU - Joraine Schmoldt
AU - Martin Wilmking
PY - 2026
DA - 2026/11/16
JO - Ecology and Evolution
DO - 10.1002/ece3.74733
UR - https://onlinelibrary.wiley.com/doi/10.1002/ece3.74733
AB - Small, self‐sufficient and affordable temperature loggers are widely used in ecological and biogeographical studies, but their accuracy, precision and possible lags relative to research grade sensors are incompletely quantified; simple, field applicable calibration procedures are needed to improve data quality and comparability. We evaluated four commonly used low‐cost loggers (HOBO U23 001A, iButton DS1925L F5#, TOMST TMS 4, TOMST Thermologger) against a Campbell Scientific EE181 L reference in a climate chamber (PERCIVAL LT 36VLX). Three replicates per logger type recorded at 5‐min intervals across six 24‐h cooling–warming cycles. The chamber was set to cycle between −10°C and +35°C; however, the sensors recorded an actual achieved range of −11.5°C to +31.5°C. Data were binned in 5°C steps to assess accuracy and precision across the gradient; we tested for time lags and derived linear, sensor specific self‐calibration formulas of the form T adj = T sensor − ( a + b × T sensor ). All loggers remained within ±1°C of the reference across the tested range, with larger errors at cold temperatures and improved performance at warmer temperatures. The EE181 L showed within‐sensor precision < ±0.1°C at ~23°C; low‐cost loggers achieved within‐sensor precision of ±0.10°C–0.15°C. All low‐cost devices exhibited short time lags of up to a few minutes in tracking temperature changes; a 5‐min timestamp shift compensated for the lag of HOBO sensors. Applying linear self‐calibration substantially reduced RMSE for all devices (e.g., iButton RMSE 0.46°C → 0.08°C; HOBO 0.22°C → 0.16°C) and improved accuracy to at least ±0.25°C across temperatures. Common low‐cost loggers provide cost effective microclimate data with high precision and adequate accuracy for many field studies but show systematic, temperature dependent offsets and short temporal lags that warrant correction. We recommend sensor and deployment specific calibrations.
SN - 2045-7758
SN - 2045-7758
ER -