Modelling diurnal and seasonal hysteresis phenomena of canopy conductance in an oasis forest ecosystem

Canopy conductance (g(c)) is considered a key and complex variable in most land surface models. Diurnal hysteretic responses have been observed for g(c) in relation to its environmental factors, posing some unavoidable challenges for the widely used g(c) models. This study combined sap flow (SF) methods with eddy covariance (EC) measurements in 2014 to investigate diurnal and seasonal variations in g(c) during a growing season of an oasis riparian Populus euphratica forest. The results showed that responses of both SF-measured g(c) and EC-measured surface conductance g(s) to environmental factors formed hysteresis loops throughout the day. Hysteresis loops of g(c) existed at different stages of growth and development in this forest and showed, significant seasonal variation. A three-step linear model and the Jarvis model were both used to simulate variations in g(c). Scalar functions of the Jarvis model showed that midday g(c) was more sensitive to D in the early growing season (from May to July), and more sensitive to R (except in May) and T in the late growing season (from August to October), while not sensitive to soil water content during the whole growing season. Although the Jarvis model effectively simulated midday g(c) the three-step linear model worked more precisely to simulate g(c) hysteresis loops throughout the growing season (r(2) = 0.88). This study highlighted the importance of dividing the daytime into different time periods and dividing the growing season into different stages to simulate diurnal and seasonal g(c) hysteresis loops. These results are helpful in achieving a better understanding of the bulk behavior of stomata in response to changing environmental conditions in oasis ecosystems.