Patch scale evapotranspiration of wetland plant species by ground-based infrared thermometry

Accurate estimation of evapotranspiration (ET) is essential to calculate the wetland water balance. However, wetlands are often composed of patches of different species of vegetation, raising the question whether different species will have different ET rates. Standard field-based meteorological methods to estimate ET, e.g., Priestley-Taylor (P-T), are sensitive to changes in climate drivers of ET but not to spatial variations in vegetation. To investigate whether ET is sensitive to vegetation type, we used thermal infrared (TIR) temperature-based surface energy balance (SEB) and Bowen ratio (β) ET estimation methods in four wetlands in northern New York State. The four wetlands had different plant communities including hardstem bulrush (Scirpus acutus), reed canary grass (Phalaris arundinacea), cattail (Typha Spp.) and meadow willow (Salix petiolaris). Across all four sites, the TIR-based methods generally had similar results as P-T methods, with mean absolute difference (MAD) of 17.1–53.0 W m⁻² and root mean squared difference (RMSD) of 23.4–62.4 W m⁻² across sites. This suggests that ET was not strongly dependent on vegetation type and topography in these northern wetlands. However, different methods showed different sensitivities to atmospheric variables including wind speed and relative humidity, suggesting that under certain conditions TIR-based methods may better represent differences in ET from different vegetation architectures.