Improving a parsimonious model for simulating urban tree transpiration using a water stress index based on within-canopy air temperature

Urban trees play a pivotal role in urban ecosystems, effectively reducing urban surface temperature by transpiration-induced evaporative cooling. However, accurately estimating tree transpiration (E_c) in urban environments is challenging due to a lack of robust methods and the inherent complexities of soil moisture monitoring in root zones, especially during dry periods. In this study, a simplified mechanistic tree transpiration model (BTA, Buckley, Turnbull & Adams) was improved by a new species-dependent normalized canopy temperature water stress index (NCTI), based on within-canopy air temperature (T_c) and reference temperature of an irrigated park (T_r). The performances of the new models (BTA-T_c) were tested using sap flow observations from two species (White Cedar: Melia azedarach; Grey Box: Eucalyptus microcarpa), with a total of seven trees in the City of Mitcham, South Australia. Results show that the BTA-T_c models reduced the overestimation of the BTA model noticeably during dry periods. Overall, the Nash–Sutcliffe efficiency (NSE) increase was 10 % for both White Cedar and Grey Box trees compared to the BTA model. Additionally, NCTI effectively captured continuous seasonal tree water stress patterns, determined by the morning air warming rate for White Cedar trees and by the afternoon average temperature for the Grey Box trees. The NCTI also revealed the benefit of a Water Sensitive Urban Design device in mitigating street tree water stress during the early dry season. Given its simple data requirements, the BTA-T_c model can be easily applied to other locations, species, or climates, offering a practical approach for monitoring urban tree water use and assessing green infrastructure performance under variable climate conditions.