UAV thermal image detects genetic trait differences among populations and genotypes of Fremont cottonwood (Populus fremontii, Salicaceae)

Abstract

Many plants are becoming increasingly maladapted to their environments due to changing climate and environmental conditions. It is, therefore, important to quantitatively evaluate what species, populations, and genotypes will survive in projected climate change scenarios and the implications this can have for associated biodiversity. We evaluate unmanned aerial vehicle (UAV)‐based high‐resolution thermal images for differentiating populations and genotypes in Fremont cottonwood (Populus fremontii S. Wats.), a foundation tree species that supports high levels of biodiversity and associated processes in riparian ecosystems. Specifically, we compare UAV thermal image‐derived tree canopy temperatures among 16 different populations and 10 replicated genotypes within two of the populations of Fremont cottonwood trees sourced from a broad environmental gradient and growing together in a common garden in central Arizona, USA. The UAV image‐derived tree canopy temperatures ranged 30°C–42°C resulting in a high overall accuracy of 85% in tree canopy classification. Our results indicate that the UAV thermal image‐derived mean tree canopy temperatures were significantly different among most of the 16 populations (P < 0.001). Within a warm‐adapted Sonoran Desert population and a cooler High Plateau population, the UAV thermal image‐derived tree canopy temperatures were also significantly different among many genotypes (P < 0.001). Furthermore, the UAV thermal image‐derived tree canopy temperatures were significantly correlated with tree canopy cover (R2 = 0.73; P‐value < 0.001) and varied with locations across the garden. Our findings have important implications for characterizing intraspecific genetic diversity in long‐lived forest trees like Fremont cottonwood and inferences for understanding ecosystem processes and guiding restoration efforts. We suggest that UAV thermal images can be used to rapidly scale laboratory‐ and plot‐based genetics research up to the landscape level. Ecological restoration efforts informed by projected climate scenarios can benefit from the UAV‐based genetics findings to identify future climate‐adapted populations and genotypes for potential propagation sources.

Thinning increases forest resiliency during unprecedented drought

“…The mean canopy temperature in the thinned forest stand was significantly greater at 28.1 °C (SD = 3.6 °C) (p < 0.0001) compared to the 24.8 °C mean temperature (SD = 2.1 °C) in the non-thinned forest stand during the average precipitation year. This is consistent with previous studies that documented strong relationships between canopy cover and temperature 18 , 20 , 35 . Specifically, the restoration thinning had created many small patches of trees and individual trees that were isolated and sparse, which showed high canopy temperature values in 2018.…”

“…Remote sensing of forest canopy temperatures is critical in understanding drought impacts on ecosystem function because canopy temperature in arid and semi-arid environments reflects critical plant physiological processes, particularly canopy water content, transpiration, and C fluxes 16 – 18 . Many tree species maintain canopy temperature close to the ambient temperature via transpiration and evaporative cooling 19 , 20 , which requires readily available soil water.…”

“…In hotter, drier drought periods, less-available soil water reduces canopy transpiration and evaporative cooling, which can lead to hotter canopy temperatures and reduced C uptake 21 . Forest remote sensing at a landscape scale provides rapid, accurate, and accessible measurements of canopy temperature and canopy moisture content changes in response to regional drought 18 , 22 .…”

Addressing Gender Inequities in Forest Science and Research

“…Dr. Sankey develops novel UAV remote sensing to evaluate restoration impacts on forest structure [201], canopy fuels [202], forest ecohydrology, snowpack and resiliency to drought and climate change [203]. She is leading methodological developments in UAV remote sensing of genetic trait-based phenotype and genotype differences for trees in response to drought and climate stress [204].…”

From communities to individuals: Using remote sensing to inform and monitor woodland restoration

“…While hyperspectral data have been used to differentiate species, it has been rarely used to explore provenance differences in spectral profiles. Recently, provenance classifications have been explored using thermal imagery (Sankey et al 2021). They showed provenances differed in their canopy temperatures (range = 30-42°C) when planted in a common environment, resulting in an overall accuracy of 85% when assigning individual canopies to a provenance.…”