Ecologists nowadays are trying to get closer to measurements of the microclimate that organisms are actually experiencing. Weather stations are great, standardized sources of temperature data at 2 m in the air, yet organisms often relate more strongly to what happens much closer to the ground. This rapidly increasing interest in microclimate ecology is great and much needed, but sometimes it is important to take a step back and ask that one important question: how good are we actually now at measuring the temperatures that we care about?

Indeed, microclimate measurements are done with a wide range of temperature sensors and radiation shields, professionally-built or home-made creations, and we lacked a good insight in how different the results could be. The big question is: are the temperatures as these sensors measure them close to the temperature as a beetle would experience it at the same location? This is exactly what we set out to answer in a recently published paper in the journal Methods in Ecology and Evolution.

Let’s start with some good news: for measurements of soil temperatures, we don’t expect too many issues. What you measure should match fairly well with what’s actually happening. However, it is above-ground that the trouble starts. More precisely: when the sun is shining and especially when and where wind speeds are low. Indeed, most commonly used sensors yield large errors under direct sunlight, reaching up to a whopping 25°C.

Unfortunately this problem cannot be wholly overcome by shielding the thermometer from sunlight, as the shield itself will influence both the temperatures being measured and the accuracy of measurement. Importantly, however, when there is no direct sunlight, for example at night or in shaded environments like forests, errors turned out to be much smaller (see graph).

So, what to do if one wants to measure air temperature close to the ground? In our paper, we provide two clear suggestions:

1. Whenever possible, use the smallest temperature sensors you can find, as these will be affected much less by heat absorption. Low‐cost and unshielded ultrafine‐wire thermocouples were clearly ‘best of the test’, as they will affect the surrounding temperatures the least due to their small size.
2. In shaded environments, there are more options available, and in some circumstances the use of other logger types, particularly TMS4 dataloggers, is appropriate. The latter are an especially good choice when trade-offs for costs and practical use have to be made. These might also be your go-to solution when the measured effect sizes (i.e., the difference between your location and weather station data) are large compared to the expected errors, such as may occur when regional or elevational variation in temperature is the primary concern, or in locations where weather stations are sparsely distributed.

In short: there is no perfect way to measure microclimate temperatures, but there are definitely better or worse ways to do it. When working with such data, one should thus be *very* careful that no conclusions are made that should not be made.