Fire intensity, which is a function of both heat output and residence time, may affect community and population responses to burning. However, fire intensity can be difficult to estimate consistently across the larger scales required of many ecological studies. We compared data from two inexpensive, easily deployed methods of estimating fire intensity (pyrometers using a lacquer paint series melting at 24 temperatures from 66 °C to 704 °C and calorimeters constructed of aluminum cans with water) to more detailed data from thermocouples connected to dataloggers. We sampled fire intensity from four fires in undisturbed upland Florida vegetation and in vegetation pre-treated with logging, subcanopy felling, or mowing.

Fire intensity varied widely among and within fires. Pyrometers significantly predicted all eight datalogger-derived metrics; they best predicted peak one-minute mean and were superior to calorimeters for 7 of 8 datalogger metrics. Pyrometers were inaccurate at estimating maximum temperatures even though our oven tests and prior studies had found them to be accurate. Calorimeters were significantly associated with area under the maximum >60 ºC curve, but were damaged at high residence times, and inconsistent at low intensities. In our Florida ecosystems, pyrometers were better than calorimeters at detecting site differences in fire intensities.

Failure or inconsistent performance of dataloggers, calorimeters, and pyrometers, together with microscale variation, make landscape level measurements of fire intensity quite challenging. Dataloggers are useful at small scales or in controlled environments but not generally practical at the landscape level. Calorimeters predict a narrow range of fire intensity metrics and are prone to failure. Pyrometers are one of the most practical methods for estimating fire intensity from many points over large areas, but calibration may be needed to measure fire intensities in large and heterogeneous fires.