Temperature is a very important abiotic factor for living things, low temperatures slow down metabolism, and high ones can cause damage which is irreversible. Spot temperature readings, though, have limited use, as do readings just taken from the general atmosphere, outside where the organisms live. Therefore, a combination of readings over a period of time, and taken in microhabitats is the ideal. The latter is achievable with a thermistor thermometer and the former if this attached to some sort of datalogger. Fortunately, such combinations are now commonplace in schools (e.g. here) and so the only problem is the opportunity to leave the equipment out in the field for the requisite length of time. Security is obviously an issue and will need to be judged site by site.

As to where temperature is taken, a mini um of air above plants, amongst plants and soil would be sensible. Most Dataloggers can take three probes, some can take more, so this should be quite possible.