Integration of carbon dioxide concentration in a simplified process-based model for evapotranspiration estimation in an old-growth forest

Evapotranspiration (ET) is one of the most difficult hydrologic variables to estimate. In the context of climate change, the rise of CO2 emissions may improve photosynthesis but depress transpiration thus increasing plant water use efficiency (WUE). Therefore, it is of great importance and necessity to take elevated atmospheric CO2 concentration (Ca) as a limiting factor for improved assessment of ET and WUE toward a better ecosystem management. In this study, we developed and incorporated a Ca stress function into a modified process-based transpiration model and extended its application from transpiration to total evapotranspiration simulation. The original simplified process-based model (denoted as BTA by combining developers' last name initials), previously modified version with added soil moisture (BTA-theta) and the extended version in this study with CO2 concentration (BTA-Ca) were tested for daily ET simulation at the Dinghushan forest ecosystem in south China. The results show that eddy covariance based forest ET was well simulated by the models, while the two modified models outperformed the original one. Despite the limited improvement of accuracy, the BTA-Ca model is more superior and meaningful regarding the physical mechanisms. Moreover, we evaluated the biophysical responses of ET and GPP to the major environmental factors, which shows that ET and GPP responded similarly to environmental changes with close linkages between the two. This study proposed a Ca stress function and provides an effective way of coupling water and carbon exchange processes for ET simulations, which can be applied for forest water budget assessment under climate change.