Commercial production of litchi (Litchi chinensis Sonn.) in arid and semi-arid areas is almost exclusively cultivated with irrigation. In this study we have improved the calculation of the basal crop coefficient ( \(\:{\text{K}}_{\text{c}\text{b}}\) ) of litchi for estimating orchard transpiration following the Allen and Pereira (Irrig Sci 28(1):17–34, 2009) (A&P) approach. The original A&P approach calculates \(\:{\text{K}}_{\text{c}\text{b}}\) from fixed leaf resistance ( \(\:{\text{r}}_{l}\) ) values for specific growth stages, but in reality \(\:{\text{r}}_{l}\) varies depending on a number of factors such as genotype, environmental conditions, crop growth stages, management, etc. We show that significantly improved \(\:{\text{K}}_{\text{c}\text{b}}\) , and therefore transpiration estimation, can be obtained using variable values instead of fixed values of \(\:{\text{r}}_{l}\) in the A&P approach. The original A&P approach uses a typical leaf resistance \(\:{(\text{r}}_{\text{t}\text{y}\text{p}})\) of 100 s/m, a value derived for annual crops. This study derived a specific \(\:{\text{r}}_{\text{t}\text{y}\text{p}}\) of 55 s/m for litchi trees using measured data, whereas \(\:{\text{r}}_{l}\) was modelled as a variable by Jarvis leaf resistance model. Orchard transpiration was subsequently calculated as the product of the \(\:{\text{K}}_{\text{c}\text{b}}\) and the reference evapotranspiration \(\:\left({\text{E}\text{T}}_{\text{o}}\right)\) . A comparison of calculated and measured transpiration rates resulted in a coefficient of determination of 0.82, a normalized root mean square error of 0.12, a normalized mean absolute error of 0.10, and a Nash-Sutcliffe Efficiency of 0.64. We conclude that the use of the variable \(\:{\text{r}}_{l}\) and \(\:{\text{r}}_{\text{t}\text{y}\text{p}}\) which are specific to litchi trees gives a more accurate estimation of the transpiration of litchi trees.
