Challenges and limitations of applying the flux variance similarity (FVS) method to partition evapotranspiration in a montane cloud forest

Partitioning evapotranspiration components is crucial for an in-depth understanding of energy, water, and carbon cycles in agricultural and forest ecosystems. In this study, the Flux Variance Similarity (FVS) method, lauded for its capability to segregate eddy covariance datasets' evapotranspiration, was applied in Taiwan's Chi-Lan montane cloud forest and the Lien-Hua-Chih forest. However, we discovered a biased early peak of transpiration using the FVS method in the Chi-Lan montane cloud forest that did not align with the diurnal cycle of transpiration obtained from the Community Land Model, observed sap flow velocity, and net radiation. This bias is attributed to the rapid increase in specific humidity, caused by additional water vapor sources from valley wind. This factor violates the FVS method's assumptions and leads to an early peak in CO₂ fluxes describing the net primary production (NPP). Furthermore, the high relative humidity conditions from afternoon to evening contribute to a larger magnitude of leaf-level water use efficiency, primarily due to minimal gradients between intercellular and ambient water vapor concentrations. The early peak of net primary production and water use efficiency skew the diurnal course of estimated transpiration. Additionally, the substantial canopy evaporation in the morning and the uncertainty in water use efficiency during periods of high relative humidity contribute to the overall uncertainty in transpiration values. Consequently, the application of the FVS method in environments akin to the Chi-Lan montane cloud forest warrants caution due to the intrinsic uncertainty. Our research emphasizes the imperative to explore different evapotranspiration partitioning techniques, especially in topographies like mountainous regions where diurnal water vapor accumulation is swift and places that are consistently subjected to high relative humidity.