Partitioning evapotranspiration based on the total ecosystem conductance fractions of soil, interception, and canopy in different biomes

Partitioning evapotranspiration (ET) into soil evaporation (E-soil), canopy interception evaporation (E-ic), and transpiration (T) yields both comprehensive insight into hydrological processes and better water management, but it is challenging. This study proposes a modified ecosystem conductance-based Priestley-Taylor (MEC-PT) algorithm for ET partitioning based on the total ecosystem conductance (G(Total)) fractions of soil, interception, and canopy. Datasets from 24 flux towers around the world were used to estimate G(Total) by coupling aerodynamic conductance and surface conductance (G(s)). Results from the MEC-PT model were compared with those from an original best-fit ecosystem-level conductance (m-order) model that only partitions G(s) into soil and canopy domains. The superior performance of the MEC-PT model, with the inclusion of an intercepted contributor and ability to fill the data gap associated with the m-order during wet conditions, describes the robustness of this approach for partitioning ET. The MEC-PT model results might reflect dew formation that produces minimal E-ic volume under non-rainfall conditions with support from the diurnal temperature (DT) presence. The ratio of T to total ET (T/ET) was found highest in forest with 0.72 (+/- 0.17 of standard deviation), followed by savanna (0.57 +/- 0.11), cropland (0.48 +/- 0.10), and grassland (0.39 +/- 0.17). Also, sensitivity analysis was conducted with main input variables of the MEC-PT and m-order models over four land cover types and the whole study period of each site used in this study. The results demonstrated that, in general conditions, net radiation was the key driver controlling T/ET variations, whereas air temperature and wind speed indirectly and slightly affected T/ET. This study underlines that the inclusion of E-ic might bridge the gap in knowledge about ET and its components regarding canopy dynamics and ecosystem behaviors in the context of climate change.

Automated summary

This study introduces a modified ecosystem conductance-based Priestley-Taylor (MEC-PT) algorithm for partitioning evapotranspiration (ET), showing superior performance compared to the original model and highlighting the importance of considering canopy dynamics and ecosystem behaviors in understanding ET components. Additionally, the study finds variations in transpiration/evapotranspiration (T/ET) ratios across different land cover types, with net radiation identified as a key driver controlling these variations.