Can canopy temperature acquired from an airborne level be a tree health indicator in an urban environment?

Abstract

Nowadays, gathering information about tree health conditions in cities is necessary. Trees are essential in regulating urban microclimate and mitigating the urban heat island effect. Therefore, their health status should be crucial in urban vegetation monitoring. The growing number of new cameras, sensors and research methods allows for a broader application of thermal data in remote sensing vegetation studies. This research aimed to evaluate whether it is possible to use thermal infrared data to assess the health condition of selected species of deciduous trees in an urban environment. More specifically, the data must have a 3.6–4.9 µm spectral range, obtained during the day and the night. For this purpose, research was carried out in the city center of Warsaw (Poland) in 2020. During the airborne data acquisition, thermal data, laser scanning and RGB images were collected. Synchronously with airborne data, 617 ground references were obtained in different health condition classes (healthy, slightly poor condition, poor condition and dying) for five tree species: Acer platanoides, Acer pseudoplatanus, Aesculus hippocastanum, Tilia cordata and Tilia × euchlora. The results were as follows: (i) healthy trees were cooler than trees in poor condition and dying both during the daytime and nighttime; (ii) the difference in the canopy temperatures between healthy and dying trees was 1.06 °C of mean value on the nighttime data and 3.28 °C of mean value on the daytime data; (iii) all condition classes significantly differ from each other on daytime thermal data. The aerial thermal data can be considered a new alternative to hyperspectral data. Thermal sensing represents another method of assessing the health condition of trees in an urban environment – especially data obtained during the day, which can differentiate condition classes better than data obtained at night. The method based on thermal infrared and laser scanning data fusion could be a quick and efficient solution for identifying trees in poor health.