Continuous, long-term, high-frequency thermal imaging of vegetation: Uncertainties and recommended best practices

Abstract

a b s t r a c tLeaf temperature is an elementary driver of plant physiology, ecology and ecosystem productivity. Individual leaf temperature may deviate strongly from air temperature, and may vary throughout the canopy. Measurements of leaf temperature, conducted at a high spatial and temporal resolution, can improve our understanding of leaf water loss, stomatal conductance, photosynthetic rates, phenology, and atmosphere-ecosystem exchanges. However, continuous high-resolution measurement of leaf temperature outside of a controlled environment is difficult and rarely done. Here, thermal infrared cameras are used to measure leaf temperatures. We describe two long-term field measurement sites: one in a temperature deciduous forest, and the other in a subalpine conifer forest. The considerations and constraints for deploying such cameras are discussed and the temperature errors are typically +/-1 • C or smaller ( = 0.60• C, 2 = 1.20 • C). Lastly, we compare leaf temperature by species and height at hourly to multi-seasonal timescales and show that on average, leaf temperature is warmer than air temperature in a temperate forest. Leaf temperature can be uniform or heterogeneous across a scene, depending on canopy structure, leaf habit, and meteorology. With this data, we verify that leaf temperature follows classic expectations, yet exhibits noteworthy departures that require additional study and theoretical consideration.

Tropical forest temperature thresholds for gross primary productivity

“…However, for emissivity values between 0.94 and 0.96, which is reasonable for most vegetation, the error is ~1°C (Aubrecht et al. ). Furthermore, if several objects are of similar distances from the camera, the information on relative temperature differences should be accurate compared to their absolute temperatures (Faye et al.…”

“…Aubrecht et al. () showed that camera error is <1°C at air temperatures as high as 30°C, irrespective of relative humidity. At air temperatures above 30°C and relative humidity as high as 80%, tests showed that camera error remained below 2°C.…”

“…Of these fixed parameters, emissivity appears to contribute the most error (Aubrecht et al. ). However, for emissivity values between 0.94 and 0.96, which is reasonable for most vegetation, the error is ~1°C (Aubrecht et al.…”

“…Aubrecht et al. () found a small effect of distance on an object's retrieved temperatures, although increased distance affected the minimum size of the smallest resolved object and also the homogeneity of surfaces in a region of interest. Thus, the comparison between the deciduous canopy, evergreen canopy, and bark is the most robust, whereas comparisons with flowers have the most uncertainty, as they may include other canopy elements in addition to flowers.…”

“…By contrast, site‐level thermal cameras that capture just the canopy have shown that denser canopies from temperate broad‐leaved trees are in fact warmer than sparser canopies, by up to 5°C (Leuzinger and Körner , Aubrecht et al. ), highlighting the complexity of canopy thermal regimes.…”